diff --git a/feed_weekly.xml b/feed_weekly.xml index b5148da..84e11dc 100644 --- a/feed_weekly.xml +++ b/feed_weekly.xml @@ -7,741 +7,741 @@ http://www.rssboard.org/rss-specification python-feedgen en - Mon, 19 Jan 2026 00:13:23 +0000 + Mon, 26 Jan 2026 00:13:38 +0000 - ShowUI-π: Flow-based Generative Models as GUI Dexterous Hands - https://arxiv.org/abs/2512.24965 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2512.24965.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Siyuan Hu, Kevin Qinghong Lin, Mike Zheng Shou</p><p><b>Upvotes:</b> 39</p><p><b>Summary:</b> Building intelligent agents capable of dexterous manipulation is essential for achieving human-like automation in both robotics and digital environments. However, existing GUI agents rely on discrete click predictions (x,y), which prohibits free-form, closed-loop trajectories (e.g. dragging a progress bar) that require continuous, on-the-fly perception and adjustment. In this work, we develop ShowUI-π, the first flow-based generative model as GUI dexterous hand, featuring the following designs: (i) Unified Discrete-Continuous Actions, integrating discrete clicks and continuous drags within a shared model, enabling flexible adaptation across diverse interaction modes; (ii) Flow-based Action Generation for drag modeling, which predicts incremental cursor adjustments from continuous visual observations via a lightweight action expert, ensuring smooth and stable trajectories; (iii) Drag Training data and Benchmark, where we manually collect and synthesize 20K drag trajectories across five domains (e.g. PowerPoint, Adobe Premiere Pro), and introduce ScreenDrag, a benchmark with comprehensive online and offline evaluation protocols for assessing GUI agents' drag capabilities. Our experiments show that proprietary GUI agents still struggle on ScreenDrag (e.g. Operator scores 13.27, and the best Gemini-2.5-CUA reaches 22.18). In contrast, ShowUI-π achieves 26.98 with only 450M parameters, underscoring both the difficulty of the task and the effectiveness of our approach. We hope this work advances GUI agents toward human-like dexterous control in digital world. The code is available at https://github.com/showlab/showui-pi.</p> - https://arxiv.org/abs/2512.24965 - Wed, 31 Dec 2025 16:51:14 +0000 + FinVault: Benchmarking Financial Agent Safety in Execution-Grounded Environments + https://arxiv.org/abs/2601.07853 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07853.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhi Yang, Runguo Li, Qiqi Qiang, Jiashun Wang, Fangqi Lou, Mengping Li, Dongpo Cheng, Rui Xu, Heng Lian, Shuo Zhang, Xiaolong Liang, Xiaoming Huang, Zheng Wei, Zhaowei Liu, Xin Guo, Huacan Wang, Ronghao Chen, Liwen Zhang</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Financial agents powered by large language models (LLMs) are increasingly deployed for investment analysis, risk assessment, and automated decision-making, where their abilities to plan, invoke tools, and manipulate mutable state introduce new security risks in high-stakes and highly regulated financial environments. However, existing safety evaluations largely focus on language-model-level content compliance or abstract agent settings, failing to capture execution-grounded risks arising from real operational workflows and state-changing actions. To bridge this gap, we propose FinVault, the first execution-grounded security benchmark for financial agents, comprising 31 regulatory case-driven sandbox scenarios with state-writable databases and explicit compliance constraints, together with 107 real-world vulnerabilities and 963 test cases that systematically cover prompt injection, jailbreaking, financially adapted attacks, as well as benign inputs for false-positive evaluation. Experimental results reveal that existing defense mechanisms remain ineffective in realistic financial agent settings, with average attack success rates (ASR) still reaching up to 50.0\% on state-of-the-art models and remaining non-negligible even for the most robust systems (ASR 6.7\%), highlighting the limited transferability of current safety designs and the need for stronger financial-specific defenses. Our code can be found at https://github.com/aifinlab/FinVault.</p> + https://arxiv.org/abs/2601.07853 + Fri, 09 Jan 2026 03:25:45 +0000 - DrivingGen: A Comprehensive Benchmark for Generative Video World Models in Autonomous Driving - https://arxiv.org/abs/2601.01528 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.01528.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yang Zhou, Hao Shao, Letian Wang, Zhuofan Zong, Hongsheng Li, Steven L. Waslander</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> Video generation models, as one form of world models, have emerged as one of the most exciting frontiers in AI, promising agents the ability to imagine the future by modeling the temporal evolution of complex scenes. In autonomous driving, this vision gives rise to driving world models: generative simulators that imagine ego and agent futures, enabling scalable simulation, safe testing of corner cases, and rich synthetic data generation. Yet, despite fast-growing research activity, the field lacks a rigorous benchmark to measure progress and guide priorities. Existing evaluations remain limited: generic video metrics overlook safety-critical imaging factors; trajectory plausibility is rarely quantified; temporal and agent-level consistency is neglected; and controllability with respect to ego conditioning is ignored. Moreover, current datasets fail to cover the diversity of conditions required for real-world deployment. To address these gaps, we present DrivingGen, the first comprehensive benchmark for generative driving world models. DrivingGen combines a diverse evaluation dataset curated from both driving datasets and internet-scale video sources, spanning varied weather, time of day, geographic regions, and complex maneuvers, with a suite of new metrics that jointly assess visual realism, trajectory plausibility, temporal coherence, and controllability. Benchmarking 14 state-of-the-art models reveals clear trade-offs: general models look better but break physics, while driving-specific ones capture motion realistically but lag in visual quality. DrivingGen offers a unified evaluation framework to foster reliable, controllable, and deployable driving world models, enabling scalable simulation, planning, and data-driven decision-making.</p> - https://arxiv.org/abs/2601.01528 - Sun, 04 Jan 2026 13:36:21 +0000 + More Images, More Problems? A Controlled Analysis of VLM Failure Modes + https://arxiv.org/abs/2601.07812 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07812.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Anurag Das, Adrian Bulat, Alberto Baldrati, Ioannis Maniadis Metaxas, Bernt Schiele, Georgios Tzimiropoulos, Brais Martinez</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities, yet their proficiency in understanding and reasoning over multiple images remains largely unexplored. While existing benchmarks have initiated the evaluation of multi-image models, a comprehensive analysis of their core weaknesses and their causes is still lacking. In this work, we introduce MIMIC (Multi-Image Model Insights and Challenges), a new benchmark designed to rigorously evaluate the multi-image capabilities of LVLMs. Using MIMIC, we conduct a series of diagnostic experiments that reveal pervasive issues: LVLMs often fail to aggregate information across images and struggle to track or attend to multiple concepts simultaneously. To address these failures, we propose two novel complementary remedies. On the data side, we present a procedural data-generation strategy that composes single-image annotations into rich, targeted multi-image training examples. On the optimization side, we analyze layer-wise attention patterns and derive an attention-masking scheme tailored for multi-image inputs. Experiments substantially improved cross-image aggregation, while also enhancing performance on existing multi-image benchmarks, outperforming prior state of the art across tasks. Data and code will be made available at https://github.com/anurag-198/MIMIC.</p> + https://arxiv.org/abs/2601.07812 + Mon, 12 Jan 2026 18:45:13 +0000 - VIBE: Visual Instruction Based Editor - https://arxiv.org/abs/2601.02242 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.02242.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Grigorii Alekseenko, Aleksandr Gordeev, Irina Tolstykh, Bulat Suleimanov, Vladimir Dokholyan, Georgii Fedorov, Sergey Yakubson, Aleksandra Tsybina, Mikhail Chernyshov, Maksim Kuprashevich</p><p><b>Upvotes:</b> 54</p><p><b>Summary:</b> Instruction-based image editing is among the fastest developing areas in generative AI. Over the past year, the field has reached a new level, with dozens of open-source models released alongside highly capable commercial systems. However, only a limited number of open-source approaches currently achieve real-world quality. In addition, diffusion backbones, the dominant choice for these pipelines, are often large and computationally expensive for many deployments and research settings, with widely used variants typically containing 6B to 20B parameters. This paper presents a compact, high-throughput instruction-based image editing pipeline that uses a modern 2B-parameter Qwen3-VL model to guide the editing process and the 1.6B-parameter diffusion model Sana1.5 for image generation. Our design decisions across architecture, data processing, training configuration, and evaluation target low-cost inference and strict source consistency while maintaining high quality across the major edit categories feasible at this scale. Evaluated on the ImgEdit and GEdit benchmarks, the proposed method matches or exceeds the performance of substantially heavier baselines, including models with several times as many parameters and higher inference cost, and is particularly strong on edits that require preserving the input image, such as an attribute adjustment, object removal, background edits, and targeted replacement. The model fits within 24 GB of GPU memory and generates edited images at up to 2K resolution in approximately 4 seconds on an NVIDIA H100 in BF16, without additional inference optimizations or distillation.</p> - https://arxiv.org/abs/2601.02242 - Mon, 05 Jan 2026 16:17:20 +0000 + RubricHub: A Comprehensive and Highly Discriminative Rubric Dataset via Automated Coarse-to-Fine Generation + https://arxiv.org/abs/2601.08430 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08430.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Sunzhu Li, Jiale Zhao, Miteto Wei, Huimin Ren, Yang Zhou, Jingwen Yang, Shunyu Liu, Kaike Zhang, Wei Chen</p><p><b>Upvotes:</b> 56</p><p><b>Summary:</b> Reinforcement Learning with Verifiable Rewards (RLVR) has driven substantial progress in reasoning-intensive domains like mathematics. However, optimizing open-ended generation remains challenging due to the lack of ground truth. While rubric-based evaluation offers a structured proxy for verification, existing methods suffer from scalability bottlenecks and coarse criteria, resulting in a supervision ceiling effect. To address this, we propose an automated Coarse-to-Fine Rubric Generation framework. By synergizing principle-guided synthesis, multi-model aggregation, and difficulty evolution, our approach produces comprehensive and highly discriminative criteria capable of capturing the subtle nuances. Based on this framework, we introduce RubricHub, a large-scale (sim110k) and multi-domain dataset. We validate its utility through a two-stage post-training pipeline comprising Rubric-based Rejection Sampling Fine-Tuning (RuFT) and Reinforcement Learning (RuRL). Experimental results demonstrate that RubricHub unlocks significant performance gains: our post-trained Qwen3-14B achieves state-of-the-art (SOTA) results on HealthBench (69.3), surpassing proprietary frontier models such as GPT-5. The code and data will be released soon.</p> + https://arxiv.org/abs/2601.08430 + Tue, 13 Jan 2026 10:56:39 +0000 - AnyDepth: Depth Estimation Made Easy - https://arxiv.org/abs/2601.02760 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.02760.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zeyu Ren, Zeyu Zhang, Wukai Li, Qingxiang Liu, Hao Tang</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Monocular depth estimation aims to recover the depth information of 3D scenes from 2D images. Recent work has made significant progress, but its reliance on large-scale datasets and complex decoders has limited its efficiency and generalization ability. In this paper, we propose a lightweight and data-centric framework for zero-shot monocular depth estimation. We first adopt DINOv3 as the visual encoder to obtain high-quality dense features. Secondly, to address the inherent drawbacks of the complex structure of the DPT, we design the Simple Depth Transformer (SDT), a compact transformer-based decoder. Compared to the DPT, it uses a single-path feature fusion and upsampling process to reduce the computational overhead of cross-scale feature fusion, achieving higher accuracy while reducing the number of parameters by approximately 85%-89%. Furthermore, we propose a quality-based filtering strategy to filter out harmful samples, thereby reducing dataset size while improving overall training quality. Extensive experiments on five benchmarks demonstrate that our framework surpasses the DPT in accuracy. This work highlights the importance of balancing model design and data quality for achieving efficient and generalizable zero-shot depth estimation. Code: https://github.com/AIGeeksGroup/AnyDepth. Website: https://aigeeksgroup.github.io/AnyDepth.</p> - https://arxiv.org/abs/2601.02760 - Tue, 06 Jan 2026 06:51:35 +0000 + YaPO: Learnable Sparse Activation Steering Vectors for Domain Adaptation + https://arxiv.org/abs/2601.08441 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08441.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Abdelaziz Bounhar, Rania Hossam Elmohamady Elbadry, Hadi Abdine, Preslav Nakov, Michalis Vazirgiannis, Guokan Shang</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Steering Large Language Models (LLMs) through activation interventions has emerged as a lightweight alternative to fine-tuning for alignment and personalization. Recent work on Bi-directional Preference Optimization (BiPO) shows that dense steering vectors can be learned directly from preference data in a Direct Preference Optimization (DPO) fashion, enabling control over truthfulness, hallucinations, and safety behaviors. However, dense steering vectors often entangle multiple latent factors due to neuron multi-semanticity, limiting their effectiveness and stability in fine-grained settings such as cultural alignment, where closely related values and behaviors (e.g., among Middle Eastern cultures) must be distinguished. In this paper, we propose Yet another Policy Optimization (YaPO), a reference-free method that learns sparse steering vectors in the latent space of a Sparse Autoencoder (SAE). By optimizing sparse codes, YaPO produces disentangled, interpretable, and efficient steering directions. Empirically, we show that YaPO converges faster, achieves stronger performance, and exhibits improved training stability compared to dense steering baselines. Beyond cultural alignment, YaPO generalizes to a range of alignment-related behaviors, including hallucination, wealth-seeking, jailbreak, and power-seeking. Importantly, YaPO preserves general knowledge, with no measurable degradation on MMLU. Overall, our results show that YaPO provides a general recipe for efficient, stable, and fine-grained alignment of LLMs, with broad applications to controllability and domain adaptation. The associated code and data are publicly availablehttps://github.com/MBZUAI-Paris/YaPO.</p> + https://arxiv.org/abs/2601.08441 + Tue, 13 Jan 2026 11:10:13 +0000 - MMFormalizer: Multimodal Autoformalization in the Wild - https://arxiv.org/abs/2601.03017 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.03017.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jing Xiong, Qi Han, Yunta Hsieh, Hui Shen, Huajian Xin, Chaofan Tao, Chenyang Zhao, Hengyuan Zhang, Taiqiang Wu, Zhen Zhang, Haochen Wang, Zhongwei Wan, Lingpeng Kong, Ngai Wong</p><p><b>Upvotes:</b> 102</p><p><b>Summary:</b> Autoformalization, which translates natural language mathematics into formal statements to enable machine reasoning, faces fundamental challenges in the wild due to the multimodal nature of the physical world, where physics requires inferring hidden constraints (e.g., mass or energy) from visual elements. To address this, we propose MMFormalizer, which extends autoformalization beyond text by integrating adaptive grounding with entities from real-world mathematical and physical domains. MMFormalizer recursively constructs formal propositions from perceptually grounded primitives through recursive grounding and axiom composition, with adaptive recursive termination ensuring that every abstraction is supported by visual evidence and anchored in dimensional or axiomatic grounding. We evaluate MMFormalizer on a new benchmark, PhyX-AF, comprising 115 curated samples from MathVerse, PhyX, Synthetic Geometry, and Analytic Geometry, covering diverse multimodal autoformalization tasks. Results show that frontier models such as GPT-5 and Gemini-3-Pro achieve the highest compile and semantic accuracy, with GPT-5 excelling in physical reasoning, while geometry remains the most challenging domain. Overall, MMFormalizer provides a scalable framework for unified multimodal autoformalization, bridging perception and formal reasoning. To the best of our knowledge, this is the first multimodal autoformalization method capable of handling classical mechanics (derived from the Hamiltonian), as well as relativity, quantum mechanics, and thermodynamics. More details are available on our project page: MMFormalizer.github.io</p> - https://arxiv.org/abs/2601.03017 - Tue, 06 Jan 2026 13:42:51 +0000 + Your Group-Relative Advantage Is Biased + https://arxiv.org/abs/2601.08521 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08521.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Fengkai Yang, Zherui Chen, Xiaohan Wang, Xiaodong Lu, Jiajun Chai, Guojun Yin, Wei Lin, Shuai Ma, Fuzhen Zhuang, Deqing Wang, Yaodong Yang, Jianxin Li, Yikun Ban</p><p><b>Upvotes:</b> 143</p><p><b>Summary:</b> Reinforcement Learning from Verifier Rewards (RLVR) has emerged as a widely used approach for post-training large language models on reasoning tasks, with group-based methods such as GRPO and its variants gaining broad adoption. These methods rely on group-relative advantage estimation to avoid learned critics, yet its theoretical properties remain poorly understood. In this work, we uncover a fundamental issue of group-based RL: the group-relative advantage estimator is inherently biased relative to the true (expected) advantage. We provide the first theoretical analysis showing that it systematically underestimates advantages for hard prompts and overestimates them for easy prompts, leading to imbalanced exploration and exploitation. To address this issue, we propose History-Aware Adaptive Difficulty Weighting (HA-DW), an adaptive reweighting scheme that adjusts advantage estimates based on an evolving difficulty anchor and training dynamics. Both theoretical analysis and experiments on five mathematical reasoning benchmarks demonstrate that HA-DW consistently improves performance when integrated into GRPO and its variants. Our results suggest that correcting biased advantage estimation is critical for robust and efficient RLVR training.</p> + https://arxiv.org/abs/2601.08521 + Tue, 13 Jan 2026 13:03:15 +0000 - CaricatureGS: Exaggerating 3D Gaussian Splatting Faces With Gaussian Curvature - https://arxiv.org/abs/2601.03319 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.03319.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Eldad Matmon, Amit Bracha, Noam Rotstein, Ron Kimmel</p><p><b>Upvotes:</b> 50</p><p><b>Summary:</b> A photorealistic and controllable 3D caricaturization framework for faces is introduced. We start with an intrinsic Gaussian curvature-based surface exaggeration technique, which, when coupled with texture, tends to produce over-smoothed renders. To address this, we resort to 3D Gaussian Splatting (3DGS), which has recently been shown to produce realistic free-viewpoint avatars. Given a multiview sequence, we extract a FLAME mesh, solve a curvature-weighted Poisson equation, and obtain its exaggerated form. However, directly deforming the Gaussians yields poor results, necessitating the synthesis of pseudo-ground-truth caricature images by warping each frame to its exaggerated 2D representation using local affine transformations. We then devise a training scheme that alternates real and synthesized supervision, enabling a single Gaussian collection to represent both natural and exaggerated avatars. This scheme improves fidelity, supports local edits, and allows continuous control over the intensity of the caricature. In order to achieve real-time deformations, an efficient interpolation between the original and exaggerated surfaces is introduced. We further analyze and show that it has a bounded deviation from closed-form solutions. In both quantitative and qualitative evaluations, our results outperform prior work, delivering photorealistic, geometry-controlled caricature avatars.</p> - https://arxiv.org/abs/2601.03319 - Tue, 06 Jan 2026 13:56:28 +0000 + Multiplex Thinking: Reasoning via Token-wise Branch-and-Merge + https://arxiv.org/abs/2601.08808 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08808.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yao Tang, Li Dong, Yaru Hao, Qingxiu Dong, Furu Wei, Jiatao Gu</p><p><b>Upvotes:</b> 38</p><p><b>Summary:</b> Large language models often solve complex reasoning tasks more effectively with Chain-of-Thought (CoT), but at the cost of long, low-bandwidth token sequences. Humans, by contrast, often reason softly by maintaining a distribution over plausible next steps. Motivated by this, we propose Multiplex Thinking, a stochastic soft reasoning mechanism that, at each thinking step, samples K candidate tokens and aggregates their embeddings into a single continuous multiplex token. This preserves the vocabulary embedding prior and the sampling dynamics of standard discrete generation, while inducing a tractable probability distribution over multiplex rollouts. Consequently, multiplex trajectories can be directly optimized with on-policy reinforcement learning (RL). Importantly, Multiplex Thinking is self-adaptive: when the model is confident, the multiplex token is nearly discrete and behaves like standard CoT; when it is uncertain, it compactly represents multiple plausible next steps without increasing sequence length. Across challenging math reasoning benchmarks, Multiplex Thinking consistently outperforms strong discrete CoT and RL baselines from Pass@1 through Pass@1024, while producing shorter sequences. The code and checkpoints are available at https://github.com/GMLR-Penn/Multiplex-Thinking.</p> + https://arxiv.org/abs/2601.08808 + Tue, 13 Jan 2026 18:48:00 +0000 - What Users Leave Unsaid: Under-Specified Queries Limit Vision-Language Models - https://arxiv.org/abs/2601.06165 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06165.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dasol Choi, Guijin Son, Hanwool Lee, Minhyuk Kim, Hyunwoo Ko, Teabin Lim, Ahn Eungyeol, Jungwhan Kim, Seunghyeok Hong, Youngsook Song</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Current vision-language benchmarks predominantly feature well-structured questions with clear, explicit prompts. However, real user queries are often informal and underspecified. Users naturally leave much unsaid, relying on images to convey context. We introduce HAERAE-Vision, a benchmark of 653 real-world visual questions from Korean online communities (0.76% survival from 86K candidates), each paired with an explicit rewrite, yielding 1,306 query variants in total. Evaluating 39 VLMs, we find that even state-of-the-art models (GPT-5, Gemini 2.5 Pro) achieve under 50% on the original queries. Crucially, query explicitation alone yields 8 to 22 point improvements, with smaller models benefiting most. We further show that even with web search, under-specified queries underperform explicit queries without search, revealing that current retrieval cannot compensate for what users leave unsaid. Our findings demonstrate that a substantial portion of VLM difficulty stem from natural query under-specification instead of model capability, highlighting a critical gap between benchmark evaluation and real-world deployment.</p> - https://arxiv.org/abs/2601.06165 - Wed, 07 Jan 2026 02:33:03 +0000 + Entropy Sentinel: Continuous LLM Accuracy Monitoring from Decoding Entropy Traces in STEM + https://arxiv.org/abs/2601.09001 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09001.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Pedro Memoli Buffa, Luciano Del Corro</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Deploying LLMs raises two coupled challenges: (1) monitoring - estimating where a model underperforms as traffic and domains drift - and (2) improvement - prioritizing data acquisition to close the largest performance gaps. We test whether an inference-time signal can estimate slice-level accuracy under domain shift. For each response, we compute an output-entropy profile from final-layer next-token probabilities (from top-k logprobs) and summarize it with eleven statistics. A lightweight classifier predicts instance correctness, and averaging predicted probabilities yields a domain-level accuracy estimate. We evaluate on ten STEM reasoning benchmarks with exhaustive train/test compositions (k in {1,2,3,4}; all "10 choose k" combinations), across nine LLMs from six families (3B-20B). Estimates often track held-out benchmark accuracy, and several models show near-monotonic ordering of domains. Output-entropy profiles are thus an accessible signal for scalable monitoring and for targeting data acquisition.</p> + https://arxiv.org/abs/2601.09001 + Tue, 13 Jan 2026 21:54:38 +0000 - FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selection - https://arxiv.org/abs/2601.03928 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.03928.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mingyu Ouyang, Kevin Qinghong Lin, Mike Zheng Shou, Hwee Tou Ng</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Vision-Language Models (VLMs) have shown remarkable performance in User Interface (UI) grounding tasks, driven by their ability to process increasingly high-resolution screenshots. However, screenshots are tokenized into thousands of visual tokens (e.g., about 4700 for 2K resolution), incurring significant computational overhead and diluting attention. In contrast, humans typically focus on regions of interest when interacting with UI. In this work, we pioneer the task of efficient UI grounding. Guided by practical analysis of the task's characteristics and challenges, we propose FocusUI, an efficient UI grounding framework that selects patches most relevant to the instruction while preserving positional continuity for precise grounding. FocusUI addresses two key challenges: (1) Eliminating redundant tokens in visual encoding. We construct patch-level supervision by fusing an instruction-conditioned score with a rule-based UI-graph score that down-weights large homogeneous regions to select distinct and instruction-relevant visual tokens. (2) Preserving positional continuity during visual token selection. We find that general visual token pruning methods suffer from severe accuracy degradation on UI grounding tasks due to broken positional information. We introduce a novel PosPad strategy, which compresses each contiguous sequence of dropped visual tokens into a single special marker placed at the sequence's last index to preserve positional continuity. Comprehensive experiments on four grounding benchmarks demonstrate that FocusUI surpasses GUI-specific baselines. On the ScreenSpot-Pro benchmark, FocusUI-7B achieves a performance improvement of 3.7% over GUI-Actor-7B. Even with only 30% visual token retention, FocusUI-7B drops by only 3.2% while achieving up to 1.44x faster inference and 17% lower peak GPU memory.</p> - https://arxiv.org/abs/2601.03928 - Wed, 07 Jan 2026 13:48:12 +0000 + ProFit: Leveraging High-Value Signals in SFT via Probability-Guided Token Selection + https://arxiv.org/abs/2601.09195 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09195.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tao Liu, Taiqiang Wu, Runming Yang, Shaoning Sun, Junjie Wang, Yujiu Yang</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Supervised fine-tuning (SFT) is a fundamental post-training strategy to align Large Language Models (LLMs) with human intent. However, traditional SFT often ignores the one-to-many nature of language by forcing alignment with a single reference answer, leading to the model overfitting to non-core expressions. Although our empirical analysis suggests that introducing multiple reference answers can mitigate this issue, the prohibitive data and computational costs necessitate a strategic shift: prioritizing the mitigation of single-reference overfitting over the costly pursuit of answer diversity. To achieve this, we reveal the intrinsic connection between token probability and semantic importance: high-probability tokens carry the core logical framework, while low-probability tokens are mostly replaceable expressions. Based on this insight, we propose ProFit, which selectively masks low-probability tokens to prevent surface-level overfitting. Extensive experiments confirm that ProFit consistently outperforms traditional SFT baselines on general reasoning and mathematical benchmarks.</p> + https://arxiv.org/abs/2601.09195 + Wed, 14 Jan 2026 05:50:40 +0000 - Aligning Text, Code, and Vision: A Multi-Objective Reinforcement Learning Framework for Text-to-Visualization - https://arxiv.org/abs/2601.04582 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.04582.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mizanur Rahman, Mohammed Saidul Islam, Md Tahmid Rahman Laskar, Shafiq Joty, Enamul Hoque</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Text-to-Visualization (Text2Vis) systems translate natural language queries over tabular data into concise answers and executable visualizations. While closed-source LLMs generate functional code, the resulting charts often lack semantic alignment and clarity, qualities that can only be assessed post-execution. Open-source models struggle even more, frequently producing non-executable or visually poor outputs. Although supervised fine-tuning can improve code executability, it fails to enhance overall visualization quality, as traditional SFT loss cannot capture post-execution feedback. To address this gap, we propose RL-Text2Vis, the first reinforcement learning framework for Text2Vis generation. Built on Group Relative Policy Optimization (GRPO), our method uses a novel multi-objective reward that jointly optimizes textual accuracy, code validity, and visualization quality using post-execution feedback. By training Qwen2.5 models (7B and 14B), RL-Text2Vis achieves a 22% relative improvement in chart quality over GPT-4o on the Text2Vis benchmark and boosts code execution success from 78% to 97% relative to its zero-shot baseline. Our models significantly outperform strong zero-shot and supervised baselines and also demonstrate robust generalization to out-of-domain datasets like VIS-Eval and NVBench. These results establish GRPO as an effective strategy for structured, multimodal reasoning in visualization generation. We release our code at https://github.com/vis-nlp/RL-Text2Vis.</p> - https://arxiv.org/abs/2601.04582 - Thu, 08 Jan 2026 04:29:07 +0000 + PhyRPR: Training-Free Physics-Constrained Video Generation + https://arxiv.org/abs/2601.09255 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09255.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yibo Zhao, Hengjia Li, Xiaofei He, Boxi Wu</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> Recent diffusion-based video generation models can synthesize visually plausible videos, yet they often struggle to satisfy physical constraints. A key reason is that most existing approaches remain single-stage: they entangle high-level physical understanding with low-level visual synthesis, making it hard to generate content that require explicit physical reasoning. To address this limitation, we propose a training-free three-stage pipeline,PhyRPR:Phy\uline{Reason}--Phy\uline{Plan}--Phy\uline{Refine}, which decouples physical understanding from visual synthesis. Specifically, PhyReason uses a large multimodal model for physical state reasoning and an image generator for keyframe synthesis; PhyPlan deterministically synthesizes a controllable coarse motion scaffold; and PhyRefine injects this scaffold into diffusion sampling via a latent fusion strategy to refine appearance while preserving the planned dynamics. This staged design enables explicit physical control during generation. Extensive experiments under physics constraints show that our method consistently improves physical plausibility and motion controllability.</p> + https://arxiv.org/abs/2601.09255 + Wed, 14 Jan 2026 07:41:56 +0000 - TourPlanner: A Competitive Consensus Framework with Constraint-Gated Reinforcement Learning for Travel Planning - https://arxiv.org/abs/2601.04698 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.04698.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yinuo Wang, Mining Tan, Wenxiang Jiao, Xiaoxi Li, Hao Wang, Xuanyu Zhang, Yuan Lu, Weiming Dong</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Travel planning is a sophisticated decision-making process that requires synthesizing multifaceted information to construct itineraries. However, existing travel planning approaches face several challenges: (1) Pruning candidate points of interest (POIs) while maintaining a high recall rate; (2) A single reasoning path restricts the exploration capability within the feasible solution space for travel planning; (3) Simultaneously optimizing hard constraints and soft constraints remains a significant difficulty. To address these challenges, we propose TourPlanner, a comprehensive framework featuring multi-path reasoning and constraint-gated reinforcement learning. Specifically, we first introduce a Personalized Recall and Spatial Optimization (PReSO) workflow to construct spatially-aware candidate POIs' set. Subsequently, we propose Competitive consensus Chain-of-Thought (CCoT), a multi-path reasoning paradigm that improves the ability of exploring the feasible solution space. To further refine the plan, we integrate a sigmoid-based gating mechanism into the reinforcement learning stage, which dynamically prioritizes soft-constraint satisfaction only after hard constraints are met. Experimental results on travel planning benchmarks demonstrate that TourPlanner achieves state-of-the-art performance, significantly surpassing existing methods in both feasibility and user-preference alignment.</p> - https://arxiv.org/abs/2601.04698 - Thu, 08 Jan 2026 08:08:35 +0000 + CLARE: Continual Learning for Vision-Language-Action Models via Autonomous Adapter Routing and Expansion + https://arxiv.org/abs/2601.09512 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09512.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ralf Römer, Yi Zhang, Angela P. Schoellig</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> To teach robots complex manipulation tasks, it is now a common practice to fine-tune a pre-trained vision-language-action model (VLA) on task-specific data. However, since this recipe updates existing representations, it is unsuitable for long-term operation in the real world, where robots must continually adapt to new tasks and environments while retaining the knowledge they have already acquired. Existing continual learning methods for robotics commonly require storing previous data (exemplars), struggle with long task sequences, or rely on task identifiers for deployment. To address these limitations, we propose CLARE, a general, parameter-efficient framework for exemplar-free continual learning with VLAs. CLARE introduces lightweight modular adapters into selected feedforward layers and autonomously expands the model only where necessary when learning a new task, guided by layer-wise feature similarity. During deployment, an autoencoder-based routing mechanism dynamically activates the most relevant adapters without requiring task labels. Through extensive experiments on the LIBERO benchmark, we show that CLARE achieves high performance on new tasks without catastrophic forgetting of earlier tasks, significantly outperforming even exemplar-based methods. Code and data are available at https://tum-lsy.github.io/clare.</p> + https://arxiv.org/abs/2601.09512 + Wed, 14 Jan 2026 14:23:42 +0000 - Qwen3-VL-Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the-Art Multimodal Retrieval and Ranking - https://arxiv.org/abs/2601.04720 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.04720.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mingxin Li, Yanzhao Zhang, Dingkun Long, Keqin Chen, Sibo Song, Shuai Bai, Zhibo Yang, Pengjun Xie, An Yang, Dayiheng Liu, Jingren Zhou, Junyang Lin</p><p><b>Upvotes:</b> 44</p><p><b>Summary:</b> In this report, we introduce the Qwen3-VL-Embedding and Qwen3-VL-Reranker model series, the latest extensions of the Qwen family built on the Qwen3-VL foundation model. Together, they provide an end-to-end pipeline for high-precision multimodal search by mapping diverse modalities, including text, images, document images, and video, into a unified representation space. The Qwen3-VL-Embedding model employs a multi-stage training paradigm, progressing from large-scale contrastive pre-training to reranking model distillation, to generate semantically rich high-dimensional vectors. It supports Matryoshka Representation Learning, enabling flexible embedding dimensions, and handles inputs up to 32k tokens. Complementing this, Qwen3-VL-Reranker performs fine-grained relevance estimation for query-document pairs using a cross-encoder architecture with cross-attention mechanisms. Both model series inherit the multilingual capabilities of Qwen3-VL, supporting more than 30 languages, and are released in 2B and 8B parameter sizes to accommodate diverse deployment requirements. Empirical evaluations demonstrate that the Qwen3-VL-Embedding series achieves state-of-the-art results across diverse multimodal embedding evaluation benchmarks. Specifically, Qwen3-VL-Embedding-8B attains an overall score of 77.8 on MMEB-V2, ranking first among all models (as of January 8, 2025). This report presents the architecture, training methodology, and practical capabilities of the series, demonstrating their effectiveness on various multimodal retrieval tasks, including image-text retrieval, visual question answering, and video-text matching.</p> - https://arxiv.org/abs/2601.04720 - Thu, 08 Jan 2026 08:36:06 +0000 + PersonalAlign: Hierarchical Implicit Intent Alignment for Personalized GUI Agent with Long-Term User-Centric Records + https://arxiv.org/abs/2601.09636 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09636.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yibo Lyu, Gongwei Chen, Rui Shao, Weili Guan, Liqiang Nie</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> While GUI agents have shown strong performance under explicit and completion instructions, real-world deployment requires aligning with users' more complex implicit intents. In this work, we highlight Hierarchical Implicit Intent Alignment for Personalized GUI Agent (PersonalAlign), a new agent task that requires agents to leverage long-term user records as persistent context to resolve omitted preferences in vague instructions and anticipate latent routines by user state for proactive assistance. To facilitate this study, we introduce AndroidIntent, a benchmark designed to evaluate agents' ability in resolving vague instructions and providing proactive suggestions through reasoning over long-term user records. We annotated 775 user-specific preferences and 215 routines from 20k long-term records across different users for evaluation. Furthermore, we introduce Hierarchical Intent Memory Agent (HIM-Agent), which maintains a continuously updating personal memory and hierarchically organizes user preferences and routines for personalization. Finally, we evaluate a range of GUI agents on AndroidIntent, including GPT-5, Qwen3-VL, and UI-TARS, further results show that HIM-Agent significantly improves both execution and proactive performance by 15.7% and 7.3%.</p> + https://arxiv.org/abs/2601.09636 + Wed, 14 Jan 2026 17:12:48 +0000 - KnowMe-Bench: Benchmarking Person Understanding for Lifelong Digital Companions - https://arxiv.org/abs/2601.04745 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.04745.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tingyu Wu, Zhisheng Chen, Ziyan Weng, Shuhe Wang, Chenglong Li, Shuo Zhang, Sen Hu, Silin Wu, Qizhen Lan, Huacan Wang, Ronghao Chen</p><p><b>Upvotes:</b> 53</p><p><b>Summary:</b> Existing long-horizon memory benchmarks mostly use multi-turn dialogues or synthetic user histories, which makes retrieval performance an imperfect proxy for person understanding. We present \BenchName, a publicly releasable benchmark built from long-form autobiographical narratives, where actions, context, and inner thoughts provide dense evidence for inferring stable motivations and decision principles. \BenchName~reconstructs each narrative into a flashback-aware, time-anchored stream and evaluates models with evidence-linked questions spanning factual recall, subjective state attribution, and principle-level reasoning. Across diverse narrative sources, retrieval-augmented systems mainly improve factual accuracy, while errors persist on temporally grounded explanations and higher-level inferences, highlighting the need for memory mechanisms beyond retrieval. Our data is in KnowMeBench{https://github.com/QuantaAlpha/KnowMeBench}.</p> - https://arxiv.org/abs/2601.04745 - Thu, 08 Jan 2026 09:11:33 +0000 + SIN-Bench: Tracing Native Evidence Chains in Long-Context Multimodal Scientific Interleaved Literature + https://arxiv.org/abs/2601.10108 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10108.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yiming Ren, Junjie Wang, Yuxin Meng, Yihang Shi, Zhiqiang Lin, Ruihang Chu, Yiran Xu, Ziming Li, Yunfei Zhao, Zihan Wang, Yu Qiao, Ruiming Tang, Minghao Liu, Yujiu Yang</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Evaluating whether multimodal large language models truly understand long-form scientific papers remains challenging: answer-only metrics and synthetic "Needle-In-A-Haystack" tests often reward answer matching without requiring a causal, evidence-linked reasoning trace in the document. We propose the "Fish-in-the-Ocean" (FITO) paradigm, which requires models to construct explicit cross-modal evidence chains within native scientific documents. To operationalize FITO, we build SIN-Data, a scientific interleaved corpus that preserves the native interleaving of text and figures. On top of it, we construct SIN-Bench with four progressive tasks covering evidence discovery (SIN-Find), hypothesis verification (SIN-Verify), grounded QA (SIN-QA), and evidence-anchored synthesis (SIN-Summary). We further introduce "No Evidence, No Score", scoring predictions when grounded to verifiable anchors and diagnosing evidence quality via matching, relevance, and logic. Experiments on eight MLLMs show that grounding is the primary bottleneck: Gemini-3-pro achieves the best average overall score (0.573), while GPT-5 attains the highest SIN-QA answer accuracy (0.767) but underperforms on evidence-aligned overall scores, exposing a gap between correctness and traceable support.</p> + https://arxiv.org/abs/2601.10108 + Thu, 15 Jan 2026 06:25:25 +0000 - AgentOCR: Reimagining Agent History via Optical Self-Compression - https://arxiv.org/abs/2601.04786 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.04786.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lang Feng, Fuchao Yang, Feng Chen, Xin Cheng, Haiyang Xu, Zhenglin Wan, Ming Yan, Bo An</p><p><b>Upvotes:</b> 27</p><p><b>Summary:</b> Recent advances in large language models (LLMs) enable agentic systems trained with reinforcement learning (RL) over multi-turn interaction trajectories, but practical deployment is bottlenecked by rapidly growing textual histories that inflate token budgets and memory usage. We introduce AgentOCR, a framework that exploits the superior information density of visual tokens by representing the accumulated observation-action history as a compact rendered image. To make multi-turn rollouts scalable, AgentOCR proposes segment optical caching. By decomposing history into hashable segments and maintaining a visual cache, this mechanism eliminates redundant re-rendering. Beyond fixed rendering, AgentOCR introduces agentic self-compression, where the agent actively emits a compression rate and is trained with compression-aware reward to adaptively balance task success and token efficiency. We conduct extensive experiments on challenging agentic benchmarks, ALFWorld and search-based QA. Remarkably, results demonstrate that AgentOCR preserves over 95\% of text-based agent performance while substantially reducing token consumption (>50\%), yielding consistent token and memory efficiency. Our further analysis validates a 20x rendering speedup from segment optical caching and the effective strategic balancing of self-compression.</p> - https://arxiv.org/abs/2601.04786 - Thu, 08 Jan 2026 10:10:20 +0000 + Unlocking Implicit Experience: Synthesizing Tool-Use Trajectories from Text + https://arxiv.org/abs/2601.10355 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10355.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhihao Xu, Rumei Li, Jiahuan Li, Rongxiang Weng, Jingang Wang, Xunliang Cai, Xiting Wang</p><p><b>Upvotes:</b> 38</p><p><b>Summary:</b> Enabling Large Language Models (LLMs) to effectively utilize tools in multi-turn interactions is essential for building capable autonomous agents. However, acquiring diverse and realistic multi-turn tool-use data remains a significant challenge. In this work, we propose a novel text-based paradigm. We observe that textual corpora naturally contain rich, multi-step problem-solving experiences, which can serve as an untapped, scalable, and authentic data source for multi-turn tool-use tasks. Based on this insight, we introduce GEM, a data synthesis pipeline that enables the generation and extraction of multi-turn tool-use trajectories from text corpora through a four-stage process: relevance filtering, workflow & tool extraction, trajectory grounding, and complexity refinement. To reduce the computational cost, we further train a specialized Trajectory Synthesizer via supervised fine-tuning. This model distills the complex generation pipeline into an efficient, end-to-end trajectory generator. Experiments demonstrate that our GEM-32B achieve a 16.5% improvement on the BFCL V3 Multi-turn benchmark. Our models partially surpass the performance of models trained on τ - bench (Airline and Retail) in-domain data, highlighting the superior generalization capability derived from our text-based synthesis paradigm. Notably, our Trajectory Synthesizer matches the quality of the full pipeline while significantly reducing inference latency and costs.</p> + https://arxiv.org/abs/2601.10355 + Thu, 15 Jan 2026 12:58:46 +0000 - SmartSearch: Process Reward-Guided Query Refinement for Search Agents - https://arxiv.org/abs/2601.04888 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.04888.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tongyu Wen, Guanting Dong, Zhicheng Dou</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Large language model (LLM)-based search agents have proven promising for addressing knowledge-intensive problems by incorporating information retrieval capabilities. Existing works largely focus on optimizing the reasoning paradigms of search agents, yet the quality of intermediate search queries during reasoning remains overlooked. As a result, the generated queries often remain inaccurate, leading to unexpected retrieval results and ultimately limiting search agents' overall effectiveness. To mitigate this issue, we introduce SmartSearch, a framework built upon two key mechanisms: (1) Process rewards, which provide fine-grained supervision for the quality of each intermediate search query through Dual-Level Credit Assessment. (2) Query refinement, which promotes the optimization of query generation by selectively refining low-quality search queries and regenerating subsequent search rounds based on these refinements. To enable the search agent to progressively internalize the ability to improve query quality under the guidance of process rewards, we design a three-stage curriculum learning framework. This framework guides the agent through a progression from imitation, to alignment, and ultimately to generalization. Experimental results show that SmartSearch consistently surpasses existing baselines, and additional quantitative analyses further confirm its significant gains in both search efficiency and query quality. The code is available at https://github.com/MYVAE/SmartSearch.</p> - https://arxiv.org/abs/2601.04888 - Thu, 08 Jan 2026 12:39:05 +0000 + The Assistant Axis: Situating and Stabilizing the Default Persona of Language Models + https://arxiv.org/abs/2601.10387 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10387.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Christina Lu, Jack Gallagher, Jonathan Michala, Kyle Fish, Jack Lindsey</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> Large language models can represent a variety of personas but typically default to a helpful Assistant identity cultivated during post-training. We investigate the structure of the space of model personas by extracting activation directions corresponding to diverse character archetypes. Across several different models, we find that the leading component of this persona space is an "Assistant Axis," which captures the extent to which a model is operating in its default Assistant mode. Steering towards the Assistant direction reinforces helpful and harmless behavior; steering away increases the model's tendency to identify as other entities. Moreover, steering away with more extreme values often induces a mystical, theatrical speaking style. We find this axis is also present in pre-trained models, where it primarily promotes helpful human archetypes like consultants and coaches and inhibits spiritual ones. Measuring deviations along the Assistant Axis predicts "persona drift," a phenomenon where models slip into exhibiting harmful or bizarre behaviors that are uncharacteristic of their typical persona. We find that persona drift is often driven by conversations demanding meta-reflection on the model's processes or featuring emotionally vulnerable users. We show that restricting activations to a fixed region along the Assistant Axis can stabilize model behavior in these scenarios -- and also in the face of adversarial persona-based jailbreaks. Our results suggest that post-training steers models toward a particular region of persona space but only loosely tethers them to it, motivating work on training and steering strategies that more deeply anchor models to a coherent persona.</p> + https://arxiv.org/abs/2601.10387 + Thu, 15 Jan 2026 13:40:06 +0000 - Controllable Memory Usage: Balancing Anchoring and Innovation in Long-Term Human-Agent Interaction - https://arxiv.org/abs/2601.05107 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05107.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Muzhao Tian, Zisu Huang, Xiaohua Wang, Jingwen Xu, Zhengkang Guo, Qi Qian, Yuanzhe Shen, Kaitao Song, Jiakang Yuan, Changze Lv, Xiaoqing Zheng</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> As LLM-based agents are increasingly used in long-term interactions, cumulative memory is critical for enabling personalization and maintaining stylistic consistency. However, most existing systems adopt an ``all-or-nothing'' approach to memory usage: incorporating all relevant past information can lead to Memory Anchoring, where the agent is trapped by past interactions, while excluding memory entirely results in under-utilization and the loss of important interaction history. We show that an agent's reliance on memory can be modeled as an explicit and user-controllable dimension. We first introduce a behavioral metric of memory dependence to quantify the influence of past interactions on current outputs. We then propose Steerable Memory Agent, SteeM, a framework that allows users to dynamically regulate memory reliance, ranging from a fresh-start mode that promotes innovation to a high-fidelity mode that closely follows interaction history. Experiments across different scenarios demonstrate that our approach consistently outperforms conventional prompting and rigid memory masking strategies, yielding a more nuanced and effective control for personalized human-agent collaboration.</p> - https://arxiv.org/abs/2601.05107 - Thu, 08 Jan 2026 16:54:30 +0000 + Future Optical Flow Prediction Improves Robot Control & Video Generation + https://arxiv.org/abs/2601.10781 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10781.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kanchana Ranasinghe, Honglu Zhou, Yu Fang, Luyu Yang, Le Xue, Ran Xu, Caiming Xiong, Silvio Savarese, Michael S Ryoo, Juan Carlos Niebles</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> Future motion representations, such as optical flow, offer immense value for control and generative tasks. However, forecasting generalizable spatially dense motion representations remains a key challenge, and learning such forecasting from noisy, real-world data remains relatively unexplored. We introduce FOFPred, a novel language-conditioned optical flow forecasting model featuring a unified Vision-Language Model (VLM) and Diffusion architecture. This unique combination enables strong multimodal reasoning with pixel-level generative fidelity for future motion prediction. Our model is trained on web-scale human activity data-a highly scalable but unstructured source. To extract meaningful signals from this noisy video-caption data, we employ crucial data preprocessing techniques and our unified architecture with strong image pretraining. The resulting trained model is then extended to tackle two distinct downstream tasks in control and generation. Evaluations across robotic manipulation and video generation under language-driven settings establish the cross-domain versatility of FOFPred, confirming the value of a unified VLM-Diffusion architecture and scalable learning from diverse web data for future optical flow prediction.</p> + https://arxiv.org/abs/2601.10781 + Thu, 15 Jan 2026 18:49:48 +0000 - GlimpRouter: Efficient Collaborative Inference by Glimpsing One Token of Thoughts - https://arxiv.org/abs/2601.05110 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05110.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenhao Zeng, Xuteng Zhang, Yuling Shi, Chao Hu, Yuting Chen, Beijun Shen, Xiaodong Gu</p><p><b>Upvotes:</b> 27</p><p><b>Summary:</b> Large Reasoning Models (LRMs) achieve remarkable performance by explicitly generating multi-step chains of thought, but this capability incurs substantial inference latency and computational cost. Collaborative inference offers a promising solution by selectively allocating work between lightweight and large models, yet a fundamental challenge remains: determining when a reasoning step requires the capacity of a large model or the efficiency of a small model. Existing routing strategies either rely on local token probabilities or post-hoc verification, introducing significant inference overhead. In this work, we propose a novel perspective on step-wise collaboration: the difficulty of a reasoning step can be inferred from its very first token. Inspired by the "Aha Moment" phenomenon in LRMs, we show that the entropy of the initial token serves as a strong predictor of step difficulty. Building on this insight, we introduce GlimpRouter, a training-free step-wise collaboration framework. GlimpRouter employs a lightweight model to generate only the first token of each reasoning step and routes the step to a larger model only when the initial token entropy exceeds a threshold. Experiments on multiple benchmarks demonstrate that our approach significantly reduces inference latency while preserving accuracy. For instance, GlimpRouter attains a substantial 10.7% improvement in accuracy while reducing inference latency by 25.9% compared to a standalone large model on AIME25. These results suggest a simple yet effective mechanism for reasoning: allocating computation based on a glimpse of thought rather than full-step evaluation.</p> - https://arxiv.org/abs/2601.05110 - Thu, 08 Jan 2026 16:58:07 +0000 + LIBERTy: A Causal Framework for Benchmarking Concept-Based Explanations of LLMs with Structural Counterfactuals + https://arxiv.org/abs/2601.10700 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10700.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Gilat Toker, Nitay Calderon, Ohad Amosy, Roi Reichart</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Concept-based explanations quantify how high-level concepts (e.g., gender or experience) influence model behavior, which is crucial for decision-makers in high-stakes domains. Recent work evaluates the faithfulness of such explanations by comparing them to reference causal effects estimated from counterfactuals. In practice, existing benchmarks rely on costly human-written counterfactuals that serve as an imperfect proxy. To address this, we introduce a framework for constructing datasets containing structural counterfactual pairs: LIBERTy (LLM-based Interventional Benchmark for Explainability with Reference Targets). LIBERTy is grounded in explicitly defined Structured Causal Models (SCMs) of the text generation, interventions on a concept propagate through the SCM until an LLM generates the counterfactual. We introduce three datasets (disease detection, CV screening, and workplace violence prediction) together with a new evaluation metric, order-faithfulness. Using them, we evaluate a wide range of methods across five models and identify substantial headroom for improving concept-based explanations. LIBERTy also enables systematic analysis of model sensitivity to interventions: we find that proprietary LLMs show markedly reduced sensitivity to demographic concepts, likely due to post-training mitigation. Overall, LIBERTy provides a much-needed benchmark for developing faithful explainability methods.</p> + https://arxiv.org/abs/2601.10700 + Thu, 15 Jan 2026 18:54:50 +0000 - Same Claim, Different Judgment: Benchmarking Scenario-Induced Bias in Multilingual Financial Misinformation Detection - https://arxiv.org/abs/2601.05403 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05403.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhiwei Liu, Yupen Cao, Yuechen Jiang, Mohsinul Kabir, Polydoros Giannouris, Chen Xu, Ziyang Xu, Tianlei Zhu, Tariquzzaman Faisal, Triantafillos Papadopoulos, Yan Wang, Lingfei Qian, Xueqing Peng, Zhuohan Xie, Ye Yuan, Saeed Almheiri, Abdulrazzaq Alnajjar, Mingbin Chen, Harry Stuart, Paul Thompson, Prayag Tiwari, Alejandro Lopez-Lira, Xue Liu, Jimin Huang, Sophia Ananiadou</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Large language models (LLMs) have been widely applied across various domains of finance. Since their training data are largely derived from human-authored corpora, LLMs may inherit a range of human biases. Behavioral biases can lead to instability and uncertainty in decision-making, particularly when processing financial information. However, existing research on LLM bias has mainly focused on direct questioning or simplified, general-purpose settings, with limited consideration of the complex real-world financial environments and high-risk, context-sensitive, multilingual financial misinformation detection tasks (\mfmd). In this work, we propose \mfmdscen, a comprehensive benchmark for evaluating behavioral biases of LLMs in \mfmd across diverse economic scenarios. In collaboration with financial experts, we construct three types of complex financial scenarios: (i) role- and personality-based, (ii) role- and region-based, and (iii) role-based scenarios incorporating ethnicity and religious beliefs. We further develop a multilingual financial misinformation dataset covering English, Chinese, Greek, and Bengali. By integrating these scenarios with misinformation claims, \mfmdscen enables a systematic evaluation of 22 mainstream LLMs. Our findings reveal that pronounced behavioral biases persist across both commercial and open-source models. This project will be available at https://github.com/lzw108/FMD.</p> - https://arxiv.org/abs/2601.05403 - Thu, 08 Jan 2026 22:00:32 +0000 + Advances and Frontiers of LLM-based Issue Resolution in Software Engineering: A Comprehensive Survey + https://arxiv.org/abs/2601.11655 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11655.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Caihua Li, Lianghong Guo, Yanlin Wang, Daya Guo, Wei Tao, Zhenyu Shan, Mingwei Liu, Jiachi Chen, Haoyu Song, Duyu Tang, Hongyu Zhang, Zibin Zheng</p><p><b>Upvotes:</b> 59</p><p><b>Summary:</b> Issue resolution, a complex Software Engineering (SWE) task integral to real-world development, has emerged as a compelling challenge for artificial intelligence. The establishment of benchmarks like SWE-bench revealed this task as profoundly difficult for large language models, thereby significantly accelerating the evolution of autonomous coding agents. This paper presents a systematic survey of this emerging domain. We begin by examining data construction pipelines, covering automated collection and synthesis approaches. We then provide a comprehensive analysis of methodologies, spanning training-free frameworks with their modular components to training-based techniques, including supervised fine-tuning and reinforcement learning. Subsequently, we discuss critical analyses of data quality and agent behavior, alongside practical applications. Finally, we identify key challenges and outline promising directions for future research. An open-source repository is maintained at https://github.com/DeepSoftwareAnalytics/Awesome-Issue-Resolution to serve as a dynamic resource in this field.</p> + https://arxiv.org/abs/2601.11655 + Thu, 15 Jan 2026 18:55:03 +0000 - Thinking with Map: Reinforced Parallel Map-Augmented Agent for Geolocalization - https://arxiv.org/abs/2601.05432 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05432.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yuxiang Ji, Yong Wang, Ziyu Ma, Yiming Hu, Hailang Huang, Xuecai Hu, Guanhua Chen, Liaoni Wu, Xiangxiang Chu</p><p><b>Upvotes:</b> 159</p><p><b>Summary:</b> The image geolocalization task aims to predict the location where an image was taken anywhere on Earth using visual clues. Existing large vision-language model (LVLM) approaches leverage world knowledge, chain-of-thought reasoning, and agentic capabilities, but overlook a common strategy used by humans -- using maps. In this work, we first equip the model Thinking with Map ability and formulate it as an agent-in-the-map loop. We develop a two-stage optimization scheme for it, including agentic reinforcement learning (RL) followed by parallel test-time scaling (TTS). The RL strengthens the agentic capability of model to improve sampling efficiency, and the parallel TTS enables the model to explore multiple candidate paths before making the final prediction, which is crucial for geolocalization. To evaluate our method on up-to-date and in-the-wild images, we further present MAPBench, a comprehensive geolocalization training and evaluation benchmark composed entirely of real-world images. Experimental results show that our method outperforms existing open- and closed-source models on most metrics, specifically improving Acc@500m from 8.0\% to 22.1\% compared to Gemini-3-Pro with Google Search/Map grounded mode.</p> - https://arxiv.org/abs/2601.05432 - Thu, 08 Jan 2026 23:47:30 +0000 + Reasoning Models Generate Societies of Thought + https://arxiv.org/abs/2601.10825 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10825.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Junsol Kim, Shiyang Lai, Nino Scherrer, Blaise Agüera y Arcas, James Evans</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Large language models have achieved remarkable capabilities across domains, yet mechanisms underlying sophisticated reasoning remain elusive. Recent reasoning models outperform comparable instruction-tuned models on complex cognitive tasks, attributed to extended computation through longer chains of thought. Here we show that enhanced reasoning emerges not from extended computation alone, but from simulating multi-agent-like interactions -- a society of thought -- which enables diversification and debate among internal cognitive perspectives characterized by distinct personality traits and domain expertise. Through quantitative analysis and mechanistic interpretability methods applied to reasoning traces, we find that reasoning models like DeepSeek-R1 and QwQ-32B exhibit much greater perspective diversity than instruction-tuned models, activating broader conflict between heterogeneous personality- and expertise-related features during reasoning. This multi-agent structure manifests in conversational behaviors, including question-answering, perspective shifts, and the reconciliation of conflicting views, and in socio-emotional roles that characterize sharp back-and-forth conversations, together accounting for the accuracy advantage in reasoning tasks. Controlled reinforcement learning experiments reveal that base models increase conversational behaviors when rewarded solely for reasoning accuracy, and fine-tuning models with conversational scaffolding accelerates reasoning improvement over base models. These findings indicate that the social organization of thought enables effective exploration of solution spaces. We suggest that reasoning models establish a computational parallel to collective intelligence in human groups, where diversity enables superior problem-solving when systematically structured, which suggests new opportunities for agent organization to harness the wisdom of crowds.</p> + https://arxiv.org/abs/2601.10825 + Thu, 15 Jan 2026 19:52:33 +0000 - Orient Anything V2: Unifying Orientation and Rotation Understanding - https://arxiv.org/abs/2601.05573 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05573.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zehan Wang, Ziang Zhang, Jiayang Xu, Jialei Wang, Tianyu Pang, Chao Du, HengShuang Zhao, Zhou Zhao</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> This work presents Orient Anything V2, an enhanced foundation model for unified understanding of object 3D orientation and rotation from single or paired images. Building upon Orient Anything V1, which defines orientation via a single unique front face, V2 extends this capability to handle objects with diverse rotational symmetries and directly estimate relative rotations. These improvements are enabled by four key innovations: 1) Scalable 3D assets synthesized by generative models, ensuring broad category coverage and balanced data distribution; 2) An efficient, model-in-the-loop annotation system that robustly identifies 0 to N valid front faces for each object; 3) A symmetry-aware, periodic distribution fitting objective that captures all plausible front-facing orientations, effectively modeling object rotational symmetry; 4) A multi-frame architecture that directly predicts relative object rotations. Extensive experiments show that Orient Anything V2 achieves state-of-the-art zero-shot performance on orientation estimation, 6DoF pose estimation, and object symmetry recognition across 11 widely used benchmarks. The model demonstrates strong generalization, significantly broadening the applicability of orientation estimation in diverse downstream tasks.</p> - https://arxiv.org/abs/2601.05573 - Fri, 09 Jan 2026 06:43:59 +0000 + Medical SAM3: A Foundation Model for Universal Prompt-Driven Medical Image Segmentation + https://arxiv.org/abs/2601.10880 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10880.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chongcong Jiang, Tianxingjian Ding, Chuhan Song, Jiachen Tu, Ziyang Yan, Yihua Shao, Zhenyi Wang, Yuzhang Shang, Tianyu Han, Yu Tian</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Promptable segmentation foundation models such as SAM3 have demonstrated strong generalization capabilities through interactive and concept-based prompting. However, their direct applicability to medical image segmentation remains limited by severe domain shifts, the absence of privileged spatial prompts, and the need to reason over complex anatomical and volumetric structures. Here we present Medical SAM3, a foundation model for universal prompt-driven medical image segmentation, obtained by fully fine-tuning SAM3 on large-scale, heterogeneous 2D and 3D medical imaging datasets with paired segmentation masks and text prompts. Through a systematic analysis of vanilla SAM3, we observe that its performance degrades substantially on medical data, with its apparent competitiveness largely relying on strong geometric priors such as ground-truth-derived bounding boxes. These findings motivate full model adaptation beyond prompt engineering alone. By fine-tuning SAM3's model parameters on 33 datasets spanning 10 medical imaging modalities, Medical SAM3 acquires robust domain-specific representations while preserving prompt-driven flexibility. Extensive experiments across organs, imaging modalities, and dimensionalities demonstrate consistent and significant performance gains, particularly in challenging scenarios characterized by semantic ambiguity, complex morphology, and long-range 3D context. Our results establish Medical SAM3 as a universal, text-guided segmentation foundation model for medical imaging and highlight the importance of holistic model adaptation for achieving robust prompt-driven segmentation under severe domain shift. Code and model will be made available at https://github.com/AIM-Research-Lab/Medical-SAM3.</p> + https://arxiv.org/abs/2601.10880 + Thu, 15 Jan 2026 22:18:14 +0000 - PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning - https://arxiv.org/abs/2601.05593 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05593.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jingcheng Hu, Yinmin Zhang, Shijie Shang, Xiaobo Yang, Yue Peng, Zhewei Huang, Hebin Zhou, Xin Wu, Jie Cheng, Fanqi Wan, Xiangwen Kong, Chengyuan Yao, Kaiwen Yan, Ailin Huang, Hongyu Zhou, Qi Han, Zheng Ge, Daxin Jiang, Xiangyu Zhang, Heung-Yeung Shum</p><p><b>Upvotes:</b> 77</p><p><b>Summary:</b> We introduce Parallel Coordinated Reasoning (PaCoRe), a training-and-inference framework designed to overcome a central limitation of contemporary language models: their inability to scale test-time compute (TTC) far beyond sequential reasoning under a fixed context window. PaCoRe departs from the traditional sequential paradigm by driving TTC through massive parallel exploration coordinated via a message-passing architecture in multiple rounds. Each round launches many parallel reasoning trajectories, compacts their findings into context-bounded messages, and synthesizes these messages to guide the next round and ultimately produce the final answer. Trained end-to-end with large-scale, outcome-based reinforcement learning, the model masters the synthesis abilities required by PaCoRe and scales to multi-million-token effective TTC without exceeding context limits. The approach yields strong improvements across diverse domains, and notably pushes reasoning beyond frontier systems in mathematics: an 8B model reaches 94.5% on HMMT 2025, surpassing GPT-5's 93.2% by scaling effective TTC to roughly two million tokens. We open-source model checkpoints, training data, and the full inference pipeline to accelerate follow-up work.</p> - https://arxiv.org/abs/2601.05593 - Fri, 09 Jan 2026 07:24:43 +0000 + FrankenMotion: Part-level Human Motion Generation and Composition + https://arxiv.org/abs/2601.10909 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10909.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chuqiao Li, Xianghui Xie, Yong Cao, Andreas Geiger, Gerard Pons-Moll</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> Human motion generation from text prompts has made remarkable progress in recent years. However, existing methods primarily rely on either sequence-level or action-level descriptions due to the absence of fine-grained, part-level motion annotations. This limits their controllability over individual body parts. In this work, we construct a high-quality motion dataset with atomic, temporally-aware part-level text annotations, leveraging the reasoning capabilities of large language models (LLMs). Unlike prior datasets that either provide synchronized part captions with fixed time segments or rely solely on global sequence labels, our dataset captures asynchronous and semantically distinct part movements at fine temporal resolution. Based on this dataset, we introduce a diffusion-based part-aware motion generation framework, namely FrankenMotion, where each body part is guided by its own temporally-structured textual prompt. This is, to our knowledge, the first work to provide atomic, temporally-aware part-level motion annotations and have a model that allows motion generation with both spatial (body part) and temporal (atomic action) control. Experiments demonstrate that FrankenMotion outperforms all previous baseline models adapted and retrained for our setting, and our model can compose motions unseen during training. Our code and dataset will be publicly available upon publication.</p> + https://arxiv.org/abs/2601.10909 + Thu, 15 Jan 2026 23:50:07 +0000 - EnvScaler: Scaling Tool-Interactive Environments for LLM Agent via Programmatic Synthesis - https://arxiv.org/abs/2601.05808 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05808.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiaoshuai Song, Haofei Chang, Guanting Dong, Yutao Zhu, Zhicheng Dou, Ji-Rong Wen</p><p><b>Upvotes:</b> 35</p><p><b>Summary:</b> Large language models (LLMs) are expected to be trained to act as agents in various real-world environments, but this process relies on rich and varied tool-interaction sandboxes. However, access to real systems is often restricted; LLM-simulated environments are prone to hallucinations and inconsistencies; and manually built sandboxes are hard to scale. In this paper, we propose EnvScaler, an automated framework for scalable tool-interaction environments via programmatic synthesis. EnvScaler comprises two components. First, SkelBuilder constructs diverse environment skeletons through topic mining, logic modeling, and quality evaluation. Then, ScenGenerator generates multiple task scenarios and rule-based trajectory validation functions for each environment. With EnvScaler, we synthesize 191 environments and about 7K scenarios, and apply them to Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) for Qwen3 series models. Results on three benchmarks show that EnvScaler significantly improves LLMs' ability to solve tasks in complex environments involving multi-turn, multi-tool interactions. We release our code and data at https://github.com/RUC-NLPIR/EnvScaler.</p> - https://arxiv.org/abs/2601.05808 - Fri, 09 Jan 2026 14:32:06 +0000 + What Matters in Data Curation for Multimodal Reasoning? Insights from the DCVLR Challenge + https://arxiv.org/abs/2601.10922 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10922.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yosub Shin, Michael Buriek, Boris Sobolev, Pavel Bushuyeu, Vikas Kumar, Haoyang Xu, Samuel Watson, Igor Molybog</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> We study data curation for multimodal reasoning through the NeurIPS 2025 Data Curation for Vision-Language Reasoning (DCVLR) challenge, which isolates dataset selection by fixing the model and training protocol. Using a compact curated dataset derived primarily from Walton Multimodal Cold Start, our submission placed first in the challenge. Through post-competition ablations, we show that difficulty-based example selection on an aligned base dataset is the dominant driver of performance gains. Increasing dataset size does not reliably improve mean accuracy under the fixed training recipe, but mainly reduces run-to-run variance, while commonly used diversity and synthetic augmentation heuristics provide no additional benefit and often degrade performance. These results characterize DCVLR as a saturation-regime evaluation and highlight the central role of alignment and difficulty in data-efficient multimodal reasoning.</p> + https://arxiv.org/abs/2601.10922 + Fri, 16 Jan 2026 00:50:01 +0000 - Boosting Latent Diffusion Models via Disentangled Representation Alignment - https://arxiv.org/abs/2601.05823 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05823.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> John Page, Xuesong Niu, Kai Wu, Kun Gai</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> Latent Diffusion Models (LDMs) generate high-quality images by operating in a compressed latent space, typically obtained through image tokenizers such as Variational Autoencoders (VAEs). In pursuit of a generation-friendly VAE, recent studies have explored leveraging Vision Foundation Models (VFMs) as representation alignment targets for VAEs, mirroring the approach commonly adopted for LDMs. Although this yields certain performance gains, using the same alignment target for both VAEs and LDMs overlooks their fundamentally different representational requirements. We advocate that while LDMs benefit from latents retaining high-level semantic concepts, VAEs should excel in semantic disentanglement, enabling encoding of attribute-level information in a structured way. To address this, we propose the Semantic disentangled VAE (Send-VAE), explicitly optimized for disentangled representation learning through aligning its latent space with the semantic hierarchy of pre-trained VFMs. Our approach employs a non-linear mapper network to transform VAE latents, aligning them with VFMs to bridge the gap between attribute-level disentanglement and high-level semantics, facilitating effective guidance for VAE learning. We evaluate semantic disentanglement via linear probing on attribute prediction tasks, showing strong correlation with improved generation performance. Finally, using Send-VAE, we train flow-based transformers SiTs; experiments show Send-VAE significantly speeds up training and achieves a state-of-the-art FID of 1.21 and 1.75 with and without classifier-free guidance on ImageNet 256x256.</p> - https://arxiv.org/abs/2601.05823 - Fri, 09 Jan 2026 14:54:30 +0000 + When Personalization Misleads: Understanding and Mitigating Hallucinations in Personalized LLMs + https://arxiv.org/abs/2601.11000 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11000.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhongxiang Sun, Yi Zhan, Chenglei Shen, Weijie Yu, Xiao Zhang, Ming He, Jun Xu</p><p><b>Upvotes:</b> 26</p><p><b>Summary:</b> Personalized large language models (LLMs) adapt model behavior to individual users to enhance user satisfaction, yet personalization can inadvertently distort factual reasoning. We show that when personalized LLMs face factual queries, there exists a phenomenon where the model generates answers aligned with a user's prior history rather than the objective truth, resulting in personalization-induced hallucinations that degrade factual reliability and may propagate incorrect beliefs, due to representational entanglement between personalization and factual representations. To address this issue, we propose Factuality-Preserving Personalized Steering (FPPS), a lightweight inference-time approach that mitigates personalization-induced factual distortions while preserving personalized behavior. We further introduce PFQABench, the first benchmark designed to jointly evaluate factual and personalized question answering under personalization. Experiments across multiple LLM backbones and personalization methods show that FPPS substantially improves factual accuracy while maintaining personalized performance.</p> + https://arxiv.org/abs/2601.11000 + Fri, 16 Jan 2026 05:20:10 +0000 - Goal Force: Teaching Video Models To Accomplish Physics-Conditioned Goals - https://arxiv.org/abs/2601.05848 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05848.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Nate Gillman, Yinghua Zhou, Zitian Tang, Evan Luo, Arjan Chakravarthy, Daksh Aggarwal, Michael Freeman, Charles Herrmann, Chen Sun</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> Recent advancements in video generation have enabled the development of ``world models'' capable of simulating potential futures for robotics and planning. However, specifying precise goals for these models remains a challenge; text instructions are often too abstract to capture physical nuances, while target images are frequently infeasible to specify for dynamic tasks. To address this, we introduce Goal Force, a novel framework that allows users to define goals via explicit force vectors and intermediate dynamics, mirroring how humans conceptualize physical tasks. We train a video generation model on a curated dataset of synthetic causal primitives-such as elastic collisions and falling dominos-teaching it to propagate forces through time and space. Despite being trained on simple physics data, our model exhibits remarkable zero-shot generalization to complex, real-world scenarios, including tool manipulation and multi-object causal chains. Our results suggest that by grounding video generation in fundamental physical interactions, models can emerge as implicit neural physics simulators, enabling precise, physics-aware planning without reliance on external engines. We release all datasets, code, model weights, and interactive video demos at our project page.</p> - https://arxiv.org/abs/2601.05848 - Fri, 09 Jan 2026 15:23:36 +0000 + NAACL: Noise-AwAre Verbal Confidence Calibration for LLMs in RAG Systems + https://arxiv.org/abs/2601.11004 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11004.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiayu Liu, Rui Wang, Qing Zong, Qingcheng Zeng, Tianshi Zheng, Haochen Shi, Dadi Guo, Baixuan Xu, Chunyang Li, Yangqiu Song</p><p><b>Upvotes:</b> 29</p><p><b>Summary:</b> Accurately assessing model confidence is essential for deploying large language models (LLMs) in mission-critical factual domains. While retrieval-augmented generation (RAG) is widely adopted to improve grounding, confidence calibration in RAG settings remains poorly understood. We conduct a systematic study across four benchmarks, revealing that LLMs exhibit poor calibration performance due to noisy retrieved contexts. Specifically, contradictory or irrelevant evidence tends to inflate the model's false certainty, leading to severe overconfidence. To address this, we propose NAACL Rules (Noise-AwAre Confidence CaLibration Rules) to provide a principled foundation for resolving overconfidence under noise. We further design NAACL, a noise-aware calibration framework that synthesizes supervision from about 2K HotpotQA examples guided by these rules. By performing supervised fine-tuning (SFT) with this data, NAACL equips models with intrinsic noise awareness without relying on stronger teacher models. Empirical results show that NAACL yields substantial gains, improving ECE scores by 10.9% in-domain and 8.0% out-of-domain. By bridging the gap between retrieval noise and verbal calibration, NAACL paves the way for both accurate and epistemically reliable LLMs.</p> + https://arxiv.org/abs/2601.11004 + Fri, 16 Jan 2026 05:38:25 +0000 - An Empirical Study on Preference Tuning Generalization and Diversity Under Domain Shift - https://arxiv.org/abs/2601.05882 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05882.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Constantinos Karouzos, Xingwei Tan, Nikolaos Aletras</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> Preference tuning aligns pretrained language models to human judgments of quality, helpfulness, or safety by optimizing over explicit preference signals rather than likelihood alone. Prior work has shown that preference-tuning degrades performance and reduces helpfulness when evaluated outside the training domain. However, the extent to which adaptation strategies mitigate this domain shift remains unexplored. We address this challenge by conducting a comprehensive and systematic study of alignment generalization under domain shift. We compare five popular alignment objectives and various adaptation strategies from source to target, including target-domain supervised fine-tuning and pseudo-labeling, across summarization and question-answering helpfulness tasks. Our findings reveal systematic differences in generalization across alignment objectives under domain shift. We show that adaptation strategies based on pseudo-labeling can substantially reduce domain-shift degradation</p> - https://arxiv.org/abs/2601.05882 - Fri, 09 Jan 2026 15:56:55 +0000 + BAPO: Boundary-Aware Policy Optimization for Reliable Agentic Search + https://arxiv.org/abs/2601.11037 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11037.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shiyu Liu, Yongjing Yin, Jianhao Yan, Yunbo Tang, Qinggang Zhang, Bei Li, Xin Chen, Jingang Wang, Xunliang Cai, Jinsong Su</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> RL-based agentic search enables LLMs to solve complex questions via dynamic planning and external search. While this approach significantly enhances accuracy with agent policies optimized via large-scale reinforcement learning, we identify a critical gap in reliability: these agents fail to recognize their reasoning boundaries and rarely admit ``I DON'T KNOW'' (IDK) even when evidence is insufficient or reasoning reaches its limit. The lack of reliability often leads to plausible but unreliable answers, introducing significant risks in many real-world scenarios. To this end, we propose Boundary-Aware Policy Optimization (BAPO), a novel RL framework designed to cultivate reliable boundary awareness without compromising accuracy. BAPO introduces two key components: (i) a group-based boundary-aware reward that encourages an IDK response only when the reasoning reaches its limit, and (ii) an adaptive reward modulator that strategically suspends this reward during early exploration, preventing the model from exploiting IDK as a shortcut. Extensive experiments on four benchmarks demonstrate that BAPO substantially enhances the overall reliability of agentic search.</p> + https://arxiv.org/abs/2601.11037 + Fri, 16 Jan 2026 07:06:58 +0000 - Illusions of Confidence? Diagnosing LLM Truthfulness via Neighborhood Consistency - https://arxiv.org/abs/2601.05905 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05905.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Haoming Xu, Ningyuan Zhao, Yunzhi Yao, Weihong Xu, Hongru Wang, Xinle Deng, Shumin Deng, Jeff Z. Pan, Huajun Chen, Ningyu Zhang</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> As Large Language Models (LLMs) are increasingly deployed in real-world settings, correctness alone is insufficient. Reliable deployment requires maintaining truthful beliefs under contextual perturbations. Existing evaluations largely rely on point-wise confidence like Self-Consistency, which can mask brittle belief. We show that even facts answered with perfect self-consistency can rapidly collapse under mild contextual interference. To address this gap, we propose Neighbor-Consistency Belief (NCB), a structural measure of belief robustness that evaluates response coherence across a conceptual neighborhood. To validate the efficiency of NCB, we introduce a new cognitive stress-testing protocol that probes outputs stability under contextual interference. Experiments across multiple LLMs show that the performance of high-NCB data is relatively more resistant to interference. Finally, we present Structure-Aware Training (SAT), which optimizes context-invariant belief structure and reduces long-tail knowledge brittleness by approximately 30%. Code will be available at https://github.com/zjunlp/belief.</p> - https://arxiv.org/abs/2601.05905 - Fri, 09 Jan 2026 16:23:21 +0000 + AgencyBench: Benchmarking the Frontiers of Autonomous Agents in 1M-Token Real-World Contexts + https://arxiv.org/abs/2601.11044 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11044.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Keyu Li, Junhao Shi, Yang Xiao, Mohan Jiang, Jie Sun, Yunze Wu, Shijie Xia, Xiaojie Cai, Tianze Xu, Weiye Si, Wenjie Li, Dequan Wang, Pengfei Liu</p><p><b>Upvotes:</b> 33</p><p><b>Summary:</b> Large Language Models (LLMs) based autonomous agents demonstrate multifaceted capabilities to contribute substantially to economic production. However, existing benchmarks remain focused on single agentic capability, failing to capture long-horizon real-world scenarios. Moreover, the reliance on human-in-the-loop feedback for realistic tasks creates a scalability bottleneck, hindering automated rollout collection and evaluation. To bridge this gap, we introduce AgencyBench, a comprehensive benchmark derived from daily AI usage, evaluating 6 core agentic capabilities across 32 real-world scenarios, comprising 138 tasks with specific queries, deliverables, and rubrics. These scenarios require an average of 90 tool calls, 1 million tokens, and hours of execution time to resolve. To enable automated evaluation, we employ a user simulation agent to provide iterative feedback, and a Docker sandbox to conduct visual and functional rubric-based assessment. Experiments reveal that closed-source models significantly outperform open-source models (48.4% vs 32.1%). Further analysis reveals significant disparities across models in resource efficiency, feedback-driven self-correction, and specific tool-use preferences. Finally, we investigate the impact of agentic scaffolds, observing that proprietary models demonstrate superior performance within their native ecosystems (e.g., Claude-4.5-Opus via Claude-Agent-SDK), while open-source models exhibit distinct performance peaks, suggesting potential optimization for specific execution frameworks. AgencyBench serves as a critical testbed for next-generation agents, highlighting the necessity of co-optimizing model architecture with agentic frameworks. We believe this work sheds light on the future direction of autonomous agents, and we release the full benchmark and evaluation toolkit at https://github.com/GAIR-NLP/AgencyBench.</p> + https://arxiv.org/abs/2601.11044 + Fri, 16 Jan 2026 07:22:20 +0000 - Can We Predict Before Executing Machine Learning Agents? - https://arxiv.org/abs/2601.05930 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05930.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jingsheng Zheng, Jintian Zhang, Yujie Luo, Yuren Mao, Yunjun Gao, Lun Du, Huajun Chen, Ningyu Zhang</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> Autonomous machine learning agents have revolutionized scientific discovery, yet they remain constrained by a Generate-Execute-Feedback paradigm. Previous approaches suffer from a severe Execution Bottleneck, as hypothesis evaluation relies strictly on expensive physical execution. To bypass these physical constraints, we internalize execution priors to substitute costly runtime checks with instantaneous predictive reasoning, drawing inspiration from World Models. In this work, we formalize the task of Data-centric Solution Preference and construct a comprehensive corpus of 18,438 pairwise comparisons. We demonstrate that LLMs exhibit significant predictive capabilities when primed with a Verified Data Analysis Report, achieving 61.5% accuracy and robust confidence calibration. Finally, we instantiate this framework in FOREAGENT, an agent that employs a Predict-then-Verify loop, achieving a 6x acceleration in convergence while surpassing execution-based baselines by +6%. Our code and dataset will be publicly available soon at https://github.com/zjunlp/predict-before-execute.</p> - https://arxiv.org/abs/2601.05930 - Fri, 09 Jan 2026 16:44:17 +0000 + Spurious Rewards Paradox: Mechanistically Understanding How RLVR Activates Memorization Shortcuts in LLMs + https://arxiv.org/abs/2601.11061 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11061.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lecheng Yan, Ruizhe Li, Guanhua Chen, Qing Li, Jiahui Geng, Wenxi Li, Vincent Wang, Chris Lee</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Reinforcement Learning with Verifiable Rewards (RLVR) is highly effective for enhancing LLM reasoning, yet recent evidence shows models like Qwen 2.5 achieve significant gains even with spurious or incorrect rewards. We investigate this phenomenon and identify a "Perplexity Paradox": spurious RLVR triggers a divergence where answer-token perplexity drops while prompt-side coherence degrades, suggesting the model is bypassing reasoning in favor of memorization. Using Path Patching, Logit Lens, JSD analysis, and Neural Differential Equations, we uncover a hidden Anchor-Adapter circuit that facilitates this shortcut. We localize a Functional Anchor in the middle layers (L18-20) that triggers the retrieval of memorized solutions, followed by Structural Adapters in later layers (L21+) that transform representations to accommodate the shortcut signal. Finally, we demonstrate that scaling specific MLP keys within this circuit allows for bidirectional causal steering-artificially amplifying or suppressing contamination-driven performance. Our results provide a mechanistic roadmap for identifying and mitigating data contamination in RLVR-tuned models. Code is available at https://github.com/idwts/How-RLVR-Activates-Memorization-Shortcuts.</p> + https://arxiv.org/abs/2601.11061 + Fri, 16 Jan 2026 07:55:38 +0000 - VideoAR: Autoregressive Video Generation via Next-Frame & Scale Prediction - https://arxiv.org/abs/2601.05966 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.05966.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Longbin Ji, Xiaoxiong Liu, Junyuan Shang, Shuohuan Wang, Yu Sun, Hua Wu, Haifeng Wang</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> Recent advances in video generation have been dominated by diffusion and flow-matching models, which produce high-quality results but remain computationally intensive and difficult to scale. In this work, we introduce VideoAR, the first large-scale Visual Autoregressive (VAR) framework for video generation that combines multi-scale next-frame prediction with autoregressive modeling. VideoAR disentangles spatial and temporal dependencies by integrating intra-frame VAR modeling with causal next-frame prediction, supported by a 3D multi-scale tokenizer that efficiently encodes spatio-temporal dynamics. To improve long-term consistency, we propose Multi-scale Temporal RoPE, Cross-Frame Error Correction, and Random Frame Mask, which collectively mitigate error propagation and stabilize temporal coherence. Our multi-stage pretraining pipeline progressively aligns spatial and temporal learning across increasing resolutions and durations. Empirically, VideoAR achieves new state-of-the-art results among autoregressive models, improving FVD on UCF-101 from 99.5 to 88.6 while reducing inference steps by over 10x, and reaching a VBench score of 81.74-competitive with diffusion-based models an order of magnitude larger. These results demonstrate that VideoAR narrows the performance gap between autoregressive and diffusion paradigms, offering a scalable, efficient, and temporally consistent foundation for future video generation research.</p> - https://arxiv.org/abs/2601.05966 - Fri, 09 Jan 2026 17:34:59 +0000 + ABC-Bench: Benchmarking Agentic Backend Coding in Real-World Development + https://arxiv.org/abs/2601.11077 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11077.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jie Yang, Honglin Guo, Li Ji, Jiazheng Zhou, Rui Zheng, Zhikai Lei, Shuo Zhang, Zhiheng Xi, Shichun Liu, Yuxin Wang, Bo Wang, Yining Zheng, Tao Gui, Xipeng Qiu</p><p><b>Upvotes:</b> 63</p><p><b>Summary:</b> The evolution of Large Language Models (LLMs) into autonomous agents has expanded the scope of AI coding from localized code generation to complex, repository-level, and execution-driven problem solving. However, current benchmarks predominantly evaluate code logic in static contexts, neglecting the dynamic, full-process requirements of real-world engineering, particularly in backend development which demands rigorous environment configuration and service deployment. To address this gap, we introduce ABC-Bench, a benchmark explicitly designed to evaluate agentic backend coding within a realistic, executable workflow. Using a scalable automated pipeline, we curated 224 practical tasks spanning 8 languages and 19 frameworks from open-source repositories. Distinct from previous evaluations, ABC-Bench require the agents to manage the entire development lifecycle from repository exploration to instantiating containerized services and pass the external end-to-end API tests. Our extensive evaluation reveals that even state-of-the-art models struggle to deliver reliable performance on these holistic tasks, highlighting a substantial disparity between current model capabilities and the demands of practical backend engineering. Our code is available at https://github.com/OpenMOSS/ABC-Bench.</p> + https://arxiv.org/abs/2601.11077 + Fri, 16 Jan 2026 08:23:52 +0000 - The Molecular Structure of Thought: Mapping the Topology of Long Chain-of-Thought Reasoning - https://arxiv.org/abs/2601.06002 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06002.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qiguang Chen, Yantao Du, Ziniu Li, Jinhao Liu, Songyao Duan, Jiarui Guo, Minghao Liu, Jiaheng Liu, Tong Yang, Ge Zhang, Libo Qin, Wanxiang Che, Wenhao Huang</p><p><b>Upvotes:</b> 47</p><p><b>Summary:</b> Large language models (LLMs) often fail to learn effective long chain-of-thought (Long CoT) reasoning from human or non-Long-CoT LLMs imitation. To understand this, we propose that effective and learnable Long CoT trajectories feature stable molecular-like structures in unified view, which are formed by three interaction types: Deep-Reasoning (covalent-like), Self-Reflection (hydrogen-bond-like), and Self-Exploration (van der Waals-like). Analysis of distilled trajectories reveals these structures emerge from Long CoT fine-tuning, not keyword imitation. We introduce Effective Semantic Isomers and show that only bonds promoting fast entropy convergence support stable Long CoT learning, while structural competition impairs training. Drawing on these findings, we present Mole-Syn, a distribution-transfer-graph method that guides synthesis of effective Long CoT structures, boosting performance and RL stability across benchmarks.</p> - https://arxiv.org/abs/2601.06002 - Fri, 09 Jan 2026 18:39:01 +0000 + PhysRVG: Physics-Aware Unified Reinforcement Learning for Video Generative Models + https://arxiv.org/abs/2601.11087 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11087.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qiyuan Zhang, Biao Gong, Shuai Tan, Zheng Zhang, Yujun Shen, Xing Zhu, Yuyuan Li, Kelu Yao, Chunhua Shen, Changqing Zou</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Physical principles are fundamental to realistic visual simulation, but remain a significant oversight in transformer-based video generation. This gap highlights a critical limitation in rendering rigid body motion, a core tenet of classical mechanics. While computer graphics and physics-based simulators can easily model such collisions using Newton formulas, modern pretrain-finetune paradigms discard the concept of object rigidity during pixel-level global denoising. Even perfectly correct mathematical constraints are treated as suboptimal solutions (i.e., conditions) during model optimization in post-training, fundamentally limiting the physical realism of generated videos. Motivated by these considerations, we introduce, for the first time, a physics-aware reinforcement learning paradigm for video generation models that enforces physical collision rules directly in high-dimensional spaces, ensuring the physics knowledge is strictly applied rather than treated as conditions. Subsequently, we extend this paradigm to a unified framework, termed Mimicry-Discovery Cycle (MDcycle), which allows substantial fine-tuning while fully preserving the model's ability to leverage physics-grounded feedback. To validate our approach, we construct new benchmark PhysRVGBench and perform extensive qualitative and quantitative experiments to thoroughly assess its effectiveness.</p> + https://arxiv.org/abs/2601.11087 + Fri, 16 Jan 2026 08:40:10 +0000 - Chaining the Evidence: Robust Reinforcement Learning for Deep Search Agents with Citation-Aware Rubric Rewards - https://arxiv.org/abs/2601.06021 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06021.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiajie Zhang, Xin Lv, Ling Feng, Lei Hou, Juanzi Li</p><p><b>Upvotes:</b> 39</p><p><b>Summary:</b> Reinforcement learning (RL) has emerged as a critical technique for enhancing LLM-based deep search agents. However, existing approaches primarily rely on binary outcome rewards, which fail to capture the comprehensiveness and factuality of agents' reasoning process, and often lead to undesirable behaviors such as shortcut exploitation and hallucinations. To address these limitations, we propose Citation-aware Rubric Rewards (CaRR), a fine-grained reward framework for deep search agents that emphasizes reasoning comprehensiveness, factual grounding, and evidence connectivity. CaRR decomposes complex questions into verifiable single-hop rubrics and requires agents to satisfy these rubrics by explicitly identifying hidden entities, supporting them with correct citations, and constructing complete evidence chains that link to the predicted answer. We further introduce Citation-aware Group Relative Policy Optimization (C-GRPO), which combines CaRR and outcome rewards for training robust deep search agents. Experiments show that C-GRPO consistently outperforms standard outcome-based RL baselines across multiple deep search benchmarks. Our analysis also validates that C-GRPO effectively discourages shortcut exploitation, promotes comprehensive, evidence-grounded reasoning, and exhibits strong generalization to open-ended deep research tasks. Our code and data are available at https://github.com/THUDM/CaRR.</p> - https://arxiv.org/abs/2601.06021 - Fri, 09 Jan 2026 18:57:53 +0000 + CoDance: An Unbind-Rebind Paradigm for Robust Multi-Subject Animation + https://arxiv.org/abs/2601.11096 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11096.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shuai Tan, Biao Gong, Ke Ma, Yutong Feng, Qiyuan Zhang, Yan Wang, Yujun Shen, Hengshuang Zhao</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Character image animation is gaining significant importance across various domains, driven by the demand for robust and flexible multi-subject rendering. While existing methods excel in single-person animation, they struggle to handle arbitrary subject counts, diverse character types, and spatial misalignment between the reference image and the driving poses. We attribute these limitations to an overly rigid spatial binding that forces strict pixel-wise alignment between the pose and reference, and an inability to consistently rebind motion to intended subjects. To address these challenges, we propose CoDance, a novel Unbind-Rebind framework that enables the animation of arbitrary subject counts, types, and spatial configurations conditioned on a single, potentially misaligned pose sequence. Specifically, the Unbind module employs a novel pose shift encoder to break the rigid spatial binding between the pose and the reference by introducing stochastic perturbations to both poses and their latent features, thereby compelling the model to learn a location-agnostic motion representation. To ensure precise control and subject association, we then devise a Rebind module, leveraging semantic guidance from text prompts and spatial guidance from subject masks to direct the learned motion to intended characters. Furthermore, to facilitate comprehensive evaluation, we introduce a new multi-subject CoDanceBench. Extensive experiments on CoDanceBench and existing datasets show that CoDance achieves SOTA performance, exhibiting remarkable generalization across diverse subjects and spatial layouts. The code and weights will be open-sourced.</p> + https://arxiv.org/abs/2601.11096 + Fri, 16 Jan 2026 08:53:09 +0000 - RigMo: Unifying Rig and Motion Learning for Generative Animation - https://arxiv.org/abs/2601.06378 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06378.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hao Zhang, Jiahao Luo, Bohui Wan, Yizhou Zhao, Zongrui Li, Michael Vasilkovsky, Chaoyang Wang, Jian Wang, Narendra Ahuja, Bing Zhou</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Despite significant progress in 4D generation, rig and motion, the core structural and dynamic components of animation are typically modeled as separate problems. Existing pipelines rely on ground-truth skeletons and skinning weights for motion generation and treat auto-rigging as an independent process, undermining scalability and interpretability. We present RigMo, a unified generative framework that jointly learns rig and motion directly from raw mesh sequences, without any human-provided rig annotations. RigMo encodes per-vertex deformations into two compact latent spaces: a rig latent that decodes into explicit Gaussian bones and skinning weights, and a motion latent that produces time-varying SE(3) transformations. Together, these outputs define an animatable mesh with explicit structure and coherent motion, enabling feed-forward rig and motion inference for deformable objects. Beyond unified rig-motion discovery, we introduce a Motion-DiT model operating in RigMo's latent space and demonstrate that these structure-aware latents can naturally support downstream motion generation tasks. Experiments on DeformingThings4D, Objaverse-XL, and TrueBones demonstrate that RigMo learns smooth, interpretable, and physically plausible rigs, while achieving superior reconstruction and category-level generalization compared to existing auto-rigging and deformation baselines. RigMo establishes a new paradigm for unified, structure-aware, and scalable dynamic 3D modeling.</p> - https://arxiv.org/abs/2601.06378 - Sat, 10 Jan 2026 01:26:28 +0000 + FlashLabs Chroma 1.0: A Real-Time End-to-End Spoken Dialogue Model with Personalized Voice Cloning + https://arxiv.org/abs/2601.11141 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11141.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tanyu Chen, Tairan Chen, Kai Shen, Zhenghua Bao, Zhihui Zhang, Man Yuan, Yi Shi</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> Recent end-to-end spoken dialogue systems leverage speech tokenizers and neural audio codecs to enable LLMs to operate directly on discrete speech representations. However, these models often exhibit limited speaker identity preservation, hindering personalized voice interaction. In this work, we present Chroma 1.0, the first open-source, real-time, end-to-end spoken dialogue model that achieves both low-latency interaction and high-fidelity personalized voice cloning. Chroma achieves sub-second end-to-end latency through an interleaved text-audio token schedule (1:2) that supports streaming generation, while maintaining high-quality personalized voice synthesis across multi-turn conversations. Our experimental results demonstrate that Chroma achieves a 10.96% relative improvement in speaker similarity over the human baseline, with a Real-Time Factor (RTF) of 0.43, while maintaining strong reasoning and dialogue capabilities. Our code and models are publicly available at https://github.com/FlashLabs-AI-Corp/FlashLabs-Chroma and https://huggingface.co/FlashLabs/Chroma-4B .</p> + https://arxiv.org/abs/2601.11141 + Fri, 16 Jan 2026 10:00:03 +0000 - BizFinBench.v2: A Unified Dual-Mode Bilingual Benchmark for Expert-Level Financial Capability Alignment - https://arxiv.org/abs/2601.06401 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06401.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xin Guo, Rongjunchen Zhang, Guilong Lu, Xuntao Guo, Shuai Jia, Zhi Yang, Liwen Zhang</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Large language models have undergone rapid evolution, emerging as a pivotal technology for intelligence in financial operations. However, existing benchmarks are often constrained by pitfalls such as reliance on simulated or general-purpose samples and a focus on singular, offline static scenarios. Consequently, they fail to align with the requirements for authenticity and real-time responsiveness in financial services, leading to a significant discrepancy between benchmark performance and actual operational efficacy. To address this, we introduce BizFinBench.v2, the first large-scale evaluation benchmark grounded in authentic business data from both Chinese and U.S. equity markets, integrating online assessment. We performed clustering analysis on authentic user queries from financial platforms, resulting in eight fundamental tasks and two online tasks across four core business scenarios, totaling 29,578 expert-level Q&A pairs. Experimental results demonstrate that ChatGPT-5 achieves a prominent 61.5% accuracy in main tasks, though a substantial gap relative to financial experts persists; in online tasks, DeepSeek-R1 outperforms all other commercial LLMs. Error analysis further identifies the specific capability deficiencies of existing models within practical financial business contexts. BizFinBench.v2 transcends the limitations of current benchmarks, achieving a business-level deconstruction of LLM financial capabilities and providing a precise basis for evaluating efficacy in the widespread deployment of LLMs within the financial domain. The data and code are available at https://github.com/HiThink-Research/BizFinBench.v2.</p> - https://arxiv.org/abs/2601.06401 - Sat, 10 Jan 2026 02:51:53 +0000 + Language of Thought Shapes Output Diversity in Large Language Models + https://arxiv.org/abs/2601.11227 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11227.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shaoyang Xu, Wenxuan Zhang</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Output diversity is crucial for Large Language Models as it underpins pluralism and creativity. In this work, we reveal that controlling the language used during model thinking-the language of thought-provides a novel and structural source of output diversity. Our preliminary study shows that different thinking languages occupy distinct regions in a model's thinking space. Based on this observation, we study two repeated sampling strategies under multilingual thinking-Single-Language Sampling and Mixed-Language Sampling-and conduct diversity evaluation on outputs that are controlled to be in English, regardless of the thinking language used. Across extensive experiments, we demonstrate that switching the thinking language from English to non-English languages consistently increases output diversity, with a clear and consistent positive correlation such that languages farther from English in the thinking space yield larger gains. We further show that aggregating samples across multiple thinking languages yields additional improvements through compositional effects, and that scaling sampling with linguistic heterogeneity expands the model's diversity ceiling. Finally, we show that these findings translate into practical benefits in pluralistic alignment scenarios, leading to broader coverage of cultural knowledge and value orientations in LLM outputs. Our code is publicly available at https://github.com/iNLP-Lab/Multilingual-LoT-Diversity.</p> + https://arxiv.org/abs/2601.11227 + Fri, 16 Jan 2026 12:14:16 +0000 - LSRIF: Logic-Structured Reinforcement Learning for Instruction Following - https://arxiv.org/abs/2601.06431 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06431.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qingyu Ren, Qianyu He, Jingwen Chang, Jie Zeng, Jiaqing Liang, Yanghua Xiao, Han Xia, Zeye Sun, Fei Yu</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Instruction-following is critical for large language models, but real-world instructions often contain logical structures such as sequential dependencies and conditional branching. Existing methods typically construct datasets with parallel constraints and optimize average rewards, ignoring logical dependencies and yielding noisy signals. We propose a logic-structured training framework LSRIF that explicitly models instruction logic. We first construct a dataset LSRInstruct with constraint structures such as parallel, sequential, and conditional types, and then design structure-aware rewarding method LSRIF including average aggregation for parallel structures, failure-penalty propagation for sequential structures, and selective rewards for conditional branches. Experiments show LSRIF brings significant improvements in instruction-following (in-domain and out-of-domain) and general reasoning. Analysis reveals that learning with explicit logic structures brings parameter updates in attention layers and sharpens token-level attention to constraints and logical operators.</p> - https://arxiv.org/abs/2601.06431 - Sat, 10 Jan 2026 05:11:38 +0000 + AstroReason-Bench: Evaluating Unified Agentic Planning across Heterogeneous Space Planning Problems + https://arxiv.org/abs/2601.11354 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11354.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Weiyi Wang, Xinchi Chen, Jingjing Gong, Xuanjing Huang, Xipeng Qiu</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> Recent advances in agentic Large Language Models (LLMs) have positioned them as generalist planners capable of reasoning and acting across diverse tasks. However, existing agent benchmarks largely focus on symbolic or weakly grounded environments, leaving their performance in physics-constrained real-world domains underexplored. We introduce AstroReason-Bench, a comprehensive benchmark for evaluating agentic planning in Space Planning Problems (SPP), a family of high-stakes problems with heterogeneous objectives, strict physical constraints, and long-horizon decision-making. AstroReason-Bench integrates multiple scheduling regimes, including ground station communication and agile Earth observation, and provides a unified agent-oriented interaction protocol. Evaluating on a range of state-of-the-art open- and closed-source agentic LLM systems, we find that current agents substantially underperform specialized solvers, highlighting key limitations of generalist planning under realistic constraints. AstroReason-Bench offers a challenging and diagnostic testbed for future agentic research.</p> + https://arxiv.org/abs/2601.11354 + Fri, 16 Jan 2026 15:02:41 +0000 - ArenaRL: Scaling RL for Open-Ended Agents via Tournament-based Relative Ranking - https://arxiv.org/abs/2601.06487 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06487.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qiang Zhang, Boli Chen, Fanrui Zhang, Ruixue Ding, Shihang Wang, Qiuchen Wang, Yinfeng Huang, Haonan Zhang, Rongxiang Zhu, Pengyong Wang, Ailin Ren, Xin Li, Pengjun Xie, Jiawei Liu, Ning Guo, Jingren Zhou, Zheng-Jun Zha</p><p><b>Upvotes:</b> 48</p><p><b>Summary:</b> Reinforcement learning has substantially improved the performance of LLM agents on tasks with verifiable outcomes, but it still struggles on open-ended agent tasks with vast solution spaces (e.g., complex travel planning). Due to the absence of objective ground-truth for these tasks, current RL algorithms largely rely on reward models that assign scalar scores to individual responses. We contend that such pointwise scoring suffers from an inherent discrimination collapse: the reward model struggles to distinguish subtle advantages among different trajectories, resulting in scores within a group being compressed into a narrow range. Consequently, the effective reward signal becomes dominated by noise from the reward model, leading to optimization stagnation. To address this, we propose ArenaRL, a reinforcement learning paradigm that shifts from pointwise scalar scoring to intra-group relative ranking. ArenaRL introduces a process-aware pairwise evaluation mechanism, employing multi-level rubrics to assign fine-grained relative scores to trajectories. Additionally, we construct an intra-group adversarial arena and devise a tournament-based ranking scheme to obtain stable advantage signals. Empirical results confirm that the built seeded single-elimination scheme achieves nearly equivalent advantage estimation accuracy to full pairwise comparisons with O(N^2) complexity, while operating with only O(N) complexity, striking an optimal balance between efficiency and precision. Furthermore, to address the lack of full-cycle benchmarks for open-ended agents, we build Open-Travel and Open-DeepResearch, two high-quality benchmarks featuring a comprehensive pipeline covering SFT, RL training, and multi-dimensional evaluation. Extensive experiments show that ArenaRL substantially outperforms standard RL baselines, enabling LLM agents to generate more robust solutions for complex real-world tasks.</p> - https://arxiv.org/abs/2601.06487 - Sat, 10 Jan 2026 08:43:07 +0000 + Show me the evidence: Evaluating the role of evidence and natural language explanations in AI-supported fact-checking + https://arxiv.org/abs/2601.11387 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11387.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Greta Warren, Jingyi Sun, Irina Shklovski, Isabelle Augenstein</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> Although much research has focused on AI explanations to support decisions in complex information-seeking tasks such as fact-checking, the role of evidence is surprisingly under-researched. In our study, we systematically varied explanation type, AI prediction certainty, and correctness of AI system advice for non-expert participants, who evaluated the veracity of claims and AI system predictions. Participants were provided the option of easily inspecting the underlying evidence. We found that participants consistently relied on evidence to validate AI claims across all experimental conditions. When participants were presented with natural language explanations, evidence was used less frequently although they relied on it when these explanations seemed insufficient or flawed. Qualitative data suggests that participants attempted to infer evidence source reliability, despite source identities being deliberately omitted. Our results demonstrate that evidence is a key ingredient in how people evaluate the reliability of information presented by an AI system and, in combination with natural language explanations, offers valuable support for decision-making. Further research is urgently needed to understand how evidence ought to be presented and how people engage with it in practice.</p> + https://arxiv.org/abs/2601.11387 + Fri, 16 Jan 2026 15:55:31 +0000 - BabyVision: Visual Reasoning Beyond Language - https://arxiv.org/abs/2601.06521 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06521.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Liang Chen, Weichu Xie, Yiyan Liang, Hongfeng He, Hans Zhao, Zhibo Yang, Zhiqi Huang, Haoning Wu, Haoyu Lu, Y. charles, Yiping Bao, Yuantao Fan, Guopeng Li, Haiyang Shen, Xuanzhong Chen, Wendong Xu, Shuzheng Si, Zefan Cai, Wenhao Chai, Ziqi Huang, Fangfu Liu, Tianyu Liu, Baobao Chang, Xiaobo Hu, Kaiyuan Chen, Yixin Ren, Yang Liu, Yuan Gong, Kuan Li</p><p><b>Upvotes:</b> 182</p><p><b>Summary:</b> While humans develop core visual skills long before acquiring language, contemporary Multimodal LLMs (MLLMs) still rely heavily on linguistic priors to compensate for their fragile visual understanding. We uncovered a crucial fact: state-of-the-art MLLMs consistently fail on basic visual tasks that humans, even 3-year-olds, can solve effortlessly. To systematically investigate this gap, we introduce BabyVision, a benchmark designed to assess core visual abilities independent of linguistic knowledge for MLLMs. BabyVision spans a wide range of tasks, with 388 items divided into 22 subclasses across four key categories. Empirical results and human evaluation reveal that leading MLLMs perform significantly below human baselines. Gemini3-Pro-Preview scores 49.7, lagging behind 6-year-old humans and falling well behind the average adult score of 94.1. These results show despite excelling in knowledge-heavy evaluations, current MLLMs still lack fundamental visual primitives. Progress in BabyVision represents a step toward human-level visual perception and reasoning capabilities. We also explore solving visual reasoning with generation models by proposing BabyVision-Gen and automatic evaluation toolkit. Our code and benchmark data are released at https://github.com/UniPat-AI/BabyVision for reproduction.</p> - https://arxiv.org/abs/2601.06521 - Sat, 10 Jan 2026 10:42:44 +0000 + ACoT-VLA: Action Chain-of-Thought for Vision-Language-Action Models + https://arxiv.org/abs/2601.11404 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11404.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Linqing Zhong, Yi Liu, Yifei Wei, Ziyu Xiong, Maoqing Yao, Si Liu, Guanghui Ren</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> Vision-Language-Action (VLA) models have emerged as essential generalist robot policies for diverse manipulation tasks, conventionally relying on directly translating multimodal inputs into actions via Vision-Language Model (VLM) embeddings. Recent advancements have introduced explicit intermediary reasoning, such as sub-task prediction (language) or goal image synthesis (vision), to guide action generation. However, these intermediate reasoning are often indirect and inherently limited in their capacity to convey the full, granular information required for precise action execution. Instead, we posit that the most effective form of reasoning is one that deliberates directly in the action space. We introduce Action Chain-of-Thought (ACoT), a paradigm where the reasoning process itself is formulated as a structured sequence of coarse action intents that guide the final policy. In this paper, we propose ACoT-VLA, a novel architecture that materializes the ACoT paradigm. Specifically, we introduce two complementary components: an Explicit Action Reasoner (EAR) and Implicit Action Reasoner (IAR). The former proposes coarse reference trajectories as explicit action-level reasoning steps, while the latter extracts latent action priors from internal representations of multimodal input, co-forming an ACoT that conditions the downstream action head to enable grounded policy learning. Extensive experiments in real-world and simulation environments demonstrate the superiority of our proposed method, which achieves 98.5%, 84.1%, and 47.4% on LIBERO, LIBERO-Plus and VLABench, respectively.</p> + https://arxiv.org/abs/2601.11404 + Fri, 16 Jan 2026 16:17:06 +0000 - Are LLMs Vulnerable to Preference-Undermining Attacks (PUA)? A Factorial Analysis Methodology for Diagnosing the Trade-off between Preference Alignment and Real-World Validity - https://arxiv.org/abs/2601.06596 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06596.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hongjun An, Yiliang Song, Jiangan Chen, Jiawei Shao, Chi Zhang, Xuelong Li</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Large Language Model (LLM) training often optimizes for preference alignment, rewarding outputs that are perceived as helpful and interaction-friendly. However, this preference-oriented objective can be exploited: manipulative prompts can steer responses toward user-appeasing agreement and away from truth-oriented correction. In this work, we investigate whether aligned models are vulnerable to Preference-Undermining Attacks (PUA), a class of manipulative prompting strategies designed to exploit the model's desire to please user preferences at the expense of truthfulness. We propose a diagnostic methodology that provides a finer-grained and more directive analysis than aggregate benchmark scores, using a factorial evaluation framework to decompose prompt-induced shifts into interpretable effects of system objectives (truth- vs. preference-oriented) and PUA-style dialogue factors (directive control, personal derogation, conditional approval, reality denial) within a controlled 2 times 2^4 design. Surprisingly, more advanced models are sometimes more susceptible to manipulative prompts. Beyond the dominant reality-denial factor, we observe model-specific sign reversals and interactions with PUA-style factors, suggesting tailored defenses rather than uniform robustness. These findings offer a novel, reproducible factorial evaluation methodology that provides finer-grained diagnostics for post-training processes like RLHF, enabling better trade-offs in the product iteration of LLMs by offering a more nuanced understanding of preference alignment risks and the impact of manipulative prompts.</p> - https://arxiv.org/abs/2601.06596 - Sat, 10 Jan 2026 15:16:23 +0000 + PubMed-OCR: PMC Open Access OCR Annotations + https://arxiv.org/abs/2601.11425 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11425.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hunter Heidenreich, Yosheb Getachew, Olivia Dinica, Ben Elliott</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> PubMed-OCR is an OCR-centric corpus of scientific articles derived from PubMed Central Open Access PDFs. Each page image is annotated with Google Cloud Vision and released in a compact JSON schema with word-, line-, and paragraph-level bounding boxes. The corpus spans 209.5K articles (1.5M pages; ~1.3B words) and supports layout-aware modeling, coordinate-grounded QA, and evaluation of OCR-dependent pipelines. We analyze corpus characteristics (e.g., journal coverage and detected layout features) and discuss limitations, including reliance on a single OCR engine and heuristic line reconstruction. We release the data and schema to facilitate downstream research and invite extensions.</p> + https://arxiv.org/abs/2601.11425 + Fri, 16 Jan 2026 16:44:50 +0000 - EpiCaR: Knowing What You Don't Know Matters for Better Reasoning in LLMs - https://arxiv.org/abs/2601.06786 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06786.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jewon Yeom, Jaewon Sok, Seonghyeon Park, Jeongjae Park, Taesup Kim</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Improving the reasoning abilities of large language models (LLMs) has largely relied on iterative self-training with model-generated data. While effective at boosting accuracy, existing approaches primarily reinforce successful reasoning paths, incurring a substantial calibration cost: models become overconfident and lose the ability to represent uncertainty. This failure has been characterized as a form of model collapse in alignment, where predictive distributions degenerate toward low-variance point estimates. We address this issue by reframing reasoning training as an epistemic learning problem, in which models must learn not only how to reason, but also when their reasoning should be trusted. We propose epistemically-calibrated reasoning (EpiCaR) as a training objective that jointly optimizes reasoning performance and calibration, and instantiate it within an iterative supervised fine-tuning framework using explicit self-evaluation signals. Experiments on Llama-3 and Qwen-3 families demonstrate that our approach achieves Pareto-superiority over standard baselines in both accuracy and calibration, particularly in models with sufficient reasoning capacity (e.g., 3B+). This framework generalizes effectively to OOD mathematical reasoning (GSM8K) and code generation (MBPP). Ultimately, our approach enables a 3X reduction in inference compute, matching the K=30 performance of STaR with only K=10 samples in capable models.</p> - https://arxiv.org/abs/2601.06786 - Sun, 11 Jan 2026 06:21:13 +0000 + The Poisoned Apple Effect: Strategic Manipulation of Mediated Markets via Technology Expansion of AI Agents + https://arxiv.org/abs/2601.11496 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11496.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Eilam Shapira, Roi Reichart, Moshe Tennenholtz</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> The integration of AI agents into economic markets fundamentally alters the landscape of strategic interaction. We investigate the economic implications of expanding the set of available technologies in three canonical game-theoretic settings: bargaining (resource division), negotiation (asymmetric information trade), and persuasion (strategic information transmission). We find that simply increasing the choice of AI delegates can drastically shift equilibrium payoffs and regulatory outcomes, often creating incentives for regulators to proactively develop and release technologies. Conversely, we identify a strategic phenomenon termed the "Poisoned Apple" effect: an agent may release a new technology, which neither they nor their opponent ultimately uses, solely to manipulate the regulator's choice of market design in their favor. This strategic release improves the releaser's welfare at the expense of their opponent and the regulator's fairness objectives. Our findings demonstrate that static regulatory frameworks are vulnerable to manipulation via technology expansion, necessitating dynamic market designs that adapt to the evolving landscape of AI capabilities.</p> + https://arxiv.org/abs/2601.11496 + Fri, 16 Jan 2026 18:18:03 +0000 - MemGovern: Enhancing Code Agents through Learning from Governed Human Experiences - https://arxiv.org/abs/2601.06789 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06789.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qihao Wang, Ziming Cheng, Shuo Zhang, Fan Liu, Rui Xu, Heng Lian, Kunyi Wang, Xiaoming Yu, Jianghao Yin, Sen Hu, Yue Hu, Shaolei Zhang, Yanbing Liu, Ronghao Chen, Huacan Wang</p><p><b>Upvotes:</b> 74</p><p><b>Summary:</b> While autonomous software engineering (SWE) agents are reshaping programming paradigms, they currently suffer from a "closed-world" limitation: they attempt to fix bugs from scratch or solely using local context, ignoring the immense historical human experience available on platforms like GitHub. Accessing this open-world experience is hindered by the unstructured and fragmented nature of real-world issue-tracking data. In this paper, we introduce MemGovern, a framework designed to govern and transform raw GitHub data into actionable experiential memory for agents. MemGovern employs experience governance to convert human experience into agent-friendly experience cards and introduces an agentic experience search strategy that enables logic-driven retrieval of human expertise. By producing 135K governed experience cards, MemGovern achieves a significant performance boost, improving resolution rates on the SWE-bench Verified by 4.65%. As a plug-in approach, MemGovern provides a solution for agent-friendly memory infrastructure.</p> - https://arxiv.org/abs/2601.06789 - Sun, 11 Jan 2026 06:41:26 +0000 + ShapeR: Robust Conditional 3D Shape Generation from Casual Captures + https://arxiv.org/abs/2601.11514 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11514.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yawar Siddiqui, Duncan Frost, Samir Aroudj, Armen Avetisyan, Henry Howard-Jenkins, Daniel DeTone, Pierre Moulon, Qirui Wu, Zhengqin Li, Julian Straub, Richard Newcombe, Jakob Engel</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> Recent advances in 3D shape generation have achieved impressive results, but most existing methods rely on clean, unoccluded, and well-segmented inputs. Such conditions are rarely met in real-world scenarios. We present ShapeR, a novel approach for conditional 3D object shape generation from casually captured sequences. Given an image sequence, we leverage off-the-shelf visual-inertial SLAM, 3D detection algorithms, and vision-language models to extract, for each object, a set of sparse SLAM points, posed multi-view images, and machine-generated captions. A rectified flow transformer trained to effectively condition on these modalities then generates high-fidelity metric 3D shapes. To ensure robustness to the challenges of casually captured data, we employ a range of techniques including on-the-fly compositional augmentations, a curriculum training scheme spanning object- and scene-level datasets, and strategies to handle background clutter. Additionally, we introduce a new evaluation benchmark comprising 178 in-the-wild objects across 7 real-world scenes with geometry annotations. Experiments show that ShapeR significantly outperforms existing approaches in this challenging setting, achieving an improvement of 2.7x in Chamfer distance compared to state of the art.</p> + https://arxiv.org/abs/2601.11514 + Fri, 16 Jan 2026 18:51:24 +0000 - Forest Before Trees: Latent Superposition for Efficient Visual Reasoning - https://arxiv.org/abs/2601.06803 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06803.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yubo Wang, Juntian Zhang, Yichen Wu, Yankai Lin, Nils Lukas, Yuhan Liu</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> While Chain-of-Thought empowers Large Vision-Language Models with multi-step reasoning, explicit textual rationales suffer from an information bandwidth bottleneck, where continuous visual details are discarded during discrete tokenization. Recent latent reasoning methods attempt to address this challenge, but often fall prey to premature semantic collapse due to rigid autoregressive objectives. In this paper, we propose Laser, a novel paradigm that reformulates visual deduction via Dynamic Windowed Alignment Learning (DWAL). Instead of forcing a point-wise prediction, Laser aligns the latent state with a dynamic validity window of future semantics. This mechanism enforces a "Forest-before-Trees" cognitive hierarchy, enabling the model to maintain a probabilistic superposition of global features before narrowing down to local details. Crucially, Laser maintains interpretability via decodable trajectories while stabilizing unconstrained learning via Self-Refined Superposition. Extensive experiments on 6 benchmarks demonstrate that Laser achieves state-of-the-art performance among latent reasoning methods, surpassing the strong baseline Monet by 5.03% on average. Notably, it achieves these gains with extreme efficiency, reducing inference tokens by more than 97%, while demonstrating robust generalization to out-of-distribution domains.</p> - https://arxiv.org/abs/2601.06803 - Sun, 11 Jan 2026 08:30:49 +0000 + Building Production-Ready Probes For Gemini + https://arxiv.org/abs/2601.11516 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11516.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> János Kramár, Joshua Engels, Zheng Wang, Bilal Chughtai, Rohin Shah, Neel Nanda, Arthur Conmy</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Frontier language model capabilities are improving rapidly. We thus need stronger mitigations against bad actors misusing increasingly powerful systems. Prior work has shown that activation probes may be a promising misuse mitigation technique, but we identify a key remaining challenge: probes fail to generalize under important production distribution shifts. In particular, we find that the shift from short-context to long-context inputs is difficult for existing probe architectures. We propose several new probe architecture that handle this long-context distribution shift. We evaluate these probes in the cyber-offensive domain, testing their robustness against various production-relevant shifts, including multi-turn conversations, static jailbreaks, and adaptive red teaming. Our results demonstrate that while multimax addresses context length, a combination of architecture choice and training on diverse distributions is required for broad generalization. Additionally, we show that pairing probes with prompted classifiers achieves optimal accuracy at a low cost due to the computational efficiency of probes. These findings have informed the successful deployment of misuse mitigation probes in user-facing instances of Gemini, Google's frontier language model. Finally, we find early positive results using AlphaEvolve to automate improvements in both probe architecture search and adaptive red teaming, showing that automating some AI safety research is already possible.</p> + https://arxiv.org/abs/2601.11516 + Fri, 16 Jan 2026 18:54:29 +0000 - ET-Agent: Incentivizing Effective Tool-Integrated Reasoning Agent via Behavior Calibration - https://arxiv.org/abs/2601.06860 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06860.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yifei Chen, Guanting Dong, Zhicheng Dou</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Large Language Models (LLMs) can extend their parameter knowledge limits by adopting the Tool-Integrated Reasoning (TIR) paradigm. However, existing LLM-based agent training framework often focuses on answers' accuracy, overlooking specific alignment for behavior patterns. Consequently, agent often exhibits ineffective actions during TIR tasks, such as redundant and insufficient tool calls. How to calibrate erroneous behavioral patterns when executing TIR tasks, thereby exploring effective trajectories, remains an open-ended problem. In this paper, we propose ET-Agent, a training framework for calibrating agent's tool-use behavior through two synergistic perspectives: Self-evolving Data Flywheel and Behavior Calibration Training. Specifically, we introduce a self-evolutionary data flywheel to generate enhanced data, used to fine-tune LLM to improve its exploration ability. Based on this, we implement an two-phases behavior-calibration training framework. It is designed to progressively calibrate erroneous behavioral patterns to optimal behaviors. Further in-depth experiments confirm the superiority of across multiple dimensions, including correctness, efficiency, reasoning conciseness, and tool execution accuracy. Our ET-Agent framework provides practical insights for research in the TIR field. Codes can be found in https://github.com/asilverlight/ET-Agent</p> - https://arxiv.org/abs/2601.06860 - Sun, 11 Jan 2026 11:05:26 +0000 + UniX: Unifying Autoregression and Diffusion for Chest X-Ray Understanding and Generation + https://arxiv.org/abs/2601.11522 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11522.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ruiheng Zhang, Jingfeng Yao, Huangxuan Zhao, Hao Yan, Xiao He, Lei Chen, Zhou Wei, Yong Luo, Zengmao Wang, Lefei Zhang, Dacheng Tao, Bo Du</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Despite recent progress, medical foundation models still struggle to unify visual understanding and generation, as these tasks have inherently conflicting goals: semantic abstraction versus pixel-level reconstruction. Existing approaches, typically based on parameter-shared autoregressive architectures, frequently lead to compromised performance in one or both tasks. To address this, we present UniX, a next-generation unified medical foundation model for chest X-ray understanding and generation. UniX decouples the two tasks into an autoregressive branch for understanding and a diffusion branch for high-fidelity generation. Crucially, a cross-modal self-attention mechanism is introduced to dynamically guide the generation process with understanding features. Coupled with a rigorous data cleaning pipeline and a multi-stage training strategy, this architecture enables synergistic collaboration between tasks while leveraging the strengths of diffusion models for superior generation. On two representative benchmarks, UniX achieves a 46.1% improvement in understanding performance (Micro-F1) and a 24.2% gain in generation quality (FD-RadDino), using only a quarter of the parameters of LLM-CXR. By achieving performance on par with task-specific models, our work establishes a scalable paradigm for synergistic medical image understanding and generation. Codes and models are available at https://github.com/ZrH42/UniX.</p> + https://arxiv.org/abs/2601.11522 + Fri, 16 Jan 2026 18:59:58 +0000 - Watching, Reasoning, and Searching: A Video Deep Research Benchmark on Open Web for Agentic Video Reasoning - https://arxiv.org/abs/2601.06943 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06943.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chengwen Liu, Xiaomin Yu, Zhuoyue Chang, Zhe Huang, Shuo Zhang, Heng Lian, Kunyi Wang, Rui Xu, Sen Hu, Jianheng Hou, Hao Peng, Chengwei Qin, Xiaobin Hu, Hong Peng, Ronghao Chen, Huacan Wang</p><p><b>Upvotes:</b> 202</p><p><b>Summary:</b> In real-world video question answering scenarios, videos often provide only localized visual cues, while verifiable answers are distributed across the open web; models therefore need to jointly perform cross-frame clue extraction, iterative retrieval, and multi-hop reasoning-based verification. To bridge this gap, we construct the first video deep research benchmark, VideoDR. VideoDR centers on video-conditioned open-domain video question answering, requiring cross-frame visual anchor extraction, interactive web retrieval, and multi-hop reasoning over joint video-web evidence; through rigorous human annotation and quality control, we obtain high-quality video deep research samples spanning six semantic domains. We evaluate multiple closed-source and open-source multimodal large language models under both the Workflow and Agentic paradigms, and the results show that Agentic is not consistently superior to Workflow: its gains depend on a model's ability to maintain the initial video anchors over long retrieval chains. Further analysis indicates that goal drift and long-horizon consistency are the core bottlenecks. In sum, VideoDR provides a systematic benchmark for studying video agents in open-web settings and reveals the key challenges for next-generation video deep research agents.</p> - https://arxiv.org/abs/2601.06943 - Sun, 11 Jan 2026 15:07:37 +0000 + Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces + https://arxiv.org/abs/2601.11868 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11868.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mike A. Merrill, Alexander G. Shaw, Nicholas Carlini, Boxuan Li, Harsh Raj, Ivan Bercovich, Lin Shi, Jeong Yeon Shin, Thomas Walshe, E. Kelly Buchanan, Junhong Shen, Guanghao Ye, Haowei Lin, Jason Poulos, Maoyu Wang, Marianna Nezhurina, Jenia Jitsev, Di Lu, Orfeas Menis Mastromichalakis, Zhiwei Xu, Zizhao Chen, Yue Liu, Robert Zhang, Leon Liangyu Chen, Anurag Kashyap, Jan-Lucas Uslu, Jeffrey Li, Jianbo Wu, Minghao Yan, Song Bian, Vedang Sharma, Ke Sun, Steven Dillmann, Akshay Anand, Andrew Lanpouthakoun, Bardia Koopah, Changran Hu, Etash Guha, Gabriel H. S. Dreiman, Jiacheng Zhu, Karl Krauth, Li Zhong, Niklas Muennighoff, Robert Amanfu, Shangyin Tan, Shreyas Pimpalgaonkar, Tushar Aggarwal, Xiangning Lin, Xin Lan, Xuandong Zhao, Yiqing Liang, Yuanli Wang, Zilong Wang, Changzhi Zhou, David Heineman, Hange Liu, Harsh Trivedi, John Yang, Junhong Lin, Manish Shetty, Michael Yang, Nabil Omi, Negin Raoof, Shanda Li, Terry Yue Zhuo, Wuwei Lin, Yiwei Dai, Yuxin Wang, Wenhao Chai, Shang Zhou, Dariush Wahdany, Ziyu She, Jiaming Hu, Zhikang Dong, Yuxuan Zhu, Sasha Cui, Ahson Saiyed, Arinbjörn Kolbeinsson, Jesse Hu, Christopher Michael Rytting, Ryan Marten, Yixin Wang, Alex Dimakis, Andy Konwinski, Ludwig Schmidt</p><p><b>Upvotes:</b> 26</p><p><b>Summary:</b> AI agents may soon become capable of autonomously completing valuable, long-horizon tasks in diverse domains. Current benchmarks either do not measure real-world tasks, or are not sufficiently difficult to meaningfully measure frontier models. To this end, we present Terminal-Bench 2.0: a carefully curated hard benchmark composed of 89 tasks in computer terminal environments inspired by problems from real workflows. Each task features a unique environment, human-written solution, and comprehensive tests for verification. We show that frontier models and agents score less than 65\% on the benchmark and conduct an error analysis to identify areas for model and agent improvement. We publish the dataset and evaluation harness to assist developers and researchers in future work at https://www.tbench.ai/ .</p> + https://arxiv.org/abs/2601.11868 + Sat, 17 Jan 2026 01:29:30 +0000 - X-Coder: Advancing Competitive Programming with Fully Synthetic Tasks, Solutions, and Tests - https://arxiv.org/abs/2601.06953 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06953.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jie Wu, Haoling Li, Xin Zhang, Jiani Guo, Jane Luo, Steven Liu, Yangyu Huang, Ruihang Chu, Scarlett Li, Yujiu Yang</p><p><b>Upvotes:</b> 39</p><p><b>Summary:</b> Competitive programming presents great challenges for Code LLMs due to its intensive reasoning demands and high logical complexity. However, current Code LLMs still rely heavily on real-world data, which limits their scalability. In this paper, we explore a fully synthetic approach: training Code LLMs with entirely generated tasks, solutions, and test cases, to empower code reasoning models without relying on real-world data. To support this, we leverage feature-based synthesis to propose a novel data synthesis pipeline called SynthSmith. SynthSmith shows strong potential in producing diverse and challenging tasks, along with verified solutions and tests, supporting both supervised fine-tuning and reinforcement learning. Based on the proposed synthetic SFT and RL datasets, we introduce the X-Coder model series, which achieves a notable pass rate of 62.9 avg@8 on LiveCodeBench v5 and 55.8 on v6, outperforming DeepCoder-14B-Preview and AReal-boba2-14B despite having only 7B parameters. In-depth analysis reveals that scaling laws hold on our synthetic dataset, and we explore which dimensions are more effective to scale. We further provide insights into code-centric reinforcement learning and highlight the key factors that shape performance through detailed ablations and analysis. Our findings demonstrate that scaling high-quality synthetic data and adopting staged training can greatly advance code reasoning, while mitigating reliance on real-world coding data.</p> - https://arxiv.org/abs/2601.06953 - Sun, 11 Jan 2026 15:22:33 +0000 + Agentic-R: Learning to Retrieve for Agentic Search + https://arxiv.org/abs/2601.11888 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11888.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenhan Liu, Xinyu Ma, Yutao Zhu, Yuchen Li, Daiting Shi, Dawei Yin, Zhicheng Dou</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> Agentic search has recently emerged as a powerful paradigm, where an agent interleaves multi-step reasoning with on-demand retrieval to solve complex questions. Despite its success, how to design a retriever for agentic search remains largely underexplored. Existing search agents typically rely on similarity-based retrievers, while similar passages are not always useful for final answer generation. In this paper, we propose a novel retriever training framework tailored for agentic search. Unlike retrievers designed for single-turn retrieval-augmented generation (RAG) that only rely on local passage utility, we propose to use both local query-passage relevance and global answer correctness to measure passage utility in a multi-turn agentic search. We further introduce an iterative training strategy, where the search agent and the retriever are optimized bidirectionally and iteratively. Different from RAG retrievers that are only trained once with fixed questions, our retriever is continuously improved using evolving and higher-quality queries from the agent. Extensive experiments on seven single-hop and multi-hop QA benchmarks demonstrate that our retriever, termed , consistently outperforms strong baselines across different search agents. Our codes are available at: https://github.com/8421BCD/Agentic-R.</p> + https://arxiv.org/abs/2601.11888 + Sat, 17 Jan 2026 02:59:54 +0000 - RealMem: Benchmarking LLMs in Real-World Memory-Driven Interaction - https://arxiv.org/abs/2601.06966 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.06966.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Haonan Bian, Zhiyuan Yao, Sen Hu, Zishan Xu, Shaolei Zhang, Yifu Guo, Ziliang Yang, Xueran Han, Huacan Wang, Ronghao Chen</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> As Large Language Models (LLMs) evolve from static dialogue interfaces to autonomous general agents, effective memory is paramount to ensuring long-term consistency. However, existing benchmarks primarily focus on casual conversation or task-oriented dialogue, failing to capture **"long-term project-oriented"** interactions where agents must track evolving goals. To bridge this gap, we introduce **RealMem**, the first benchmark grounded in realistic project scenarios. RealMem comprises over 2,000 cross-session dialogues across eleven scenarios, utilizing natural user queries for evaluation. We propose a synthesis pipeline that integrates Project Foundation Construction, Multi-Agent Dialogue Generation, and Memory and Schedule Management to simulate the dynamic evolution of memory. Experiments reveal that current memory systems face significant challenges in managing the long-term project states and dynamic context dependencies inherent in real-world projects. Our code and datasets are available at [https://github.com/AvatarMemory/RealMemBench](https://github.com/AvatarMemory/RealMemBench).</p> - https://arxiv.org/abs/2601.06966 - Sun, 11 Jan 2026 15:49:36 +0000 + MemoryRewardBench: Benchmarking Reward Models for Long-Term Memory Management in Large Language Models + https://arxiv.org/abs/2601.11969 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.11969.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zecheng Tang, Baibei Ji, Ruoxi Sun, Haitian Wang, WangJie You, Zhang Yijun, Wenpeng Zhu, Ji Qi, Juntao Li, Min Zhang</p><p><b>Upvotes:</b> 26</p><p><b>Summary:</b> Existing works increasingly adopt memory-centric mechanisms to process long contexts in a segment manner, and effective memory management is one of the key capabilities that enables large language models to effectively propagate information across the entire sequence. Therefore, leveraging reward models (RMs) to automatically and reliably evaluate memory quality is critical. In this work, we introduce MemoryRewardBench, the first benchmark to systematically study the ability of RMs to evaluate long-term memory management processes. MemoryRewardBench covers both long-context comprehension and long-form generation tasks, featuring 10 distinct settings with different memory management patterns, with context length ranging from 8K to 128K tokens. Evaluations on 13 cutting-edge RMs indicate a diminishing performance gap between open-source and proprietary models, with newer-generation models consistently outperforming their predecessors regardless of parameter count. We further expose the capabilities and fundamental limitations of current RMs in evaluating LLM memory management across diverse settings.</p> + https://arxiv.org/abs/2601.11969 + Sat, 17 Jan 2026 09:04:53 +0000 - Solar Open Technical Report - https://arxiv.org/abs/2601.07022 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07022.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Sungrae Park, Sanghoon Kim, Jungho Cho, Gyoungjin Gim, Dawoon Jung, Mikyoung Cha, Eunhae Choo, Taekgyu Hong, Minbyul Jeong, SeHwan Joo, Minsoo Khang, Eunwon Kim, Minjeong Kim, Sujeong Kim, Yunsu Kim, Hyeonju Lee, Seunghyun Lee, Sukyung Lee, Siyoung Park, Gyungin Shin, Inseo Song, Wonho Song, Seonghoon Yang, Seungyoun Yi, Sanghoon Yoon, Jeonghyun Ko, Seyoung Song, Keunwoo Choi, Hwalsuk Lee, Sunghun Kim, Du-Seong Chang, Kyunghyun Cho, Junsuk Choe, Hwaran Lee, Jae-Gil Lee, KyungTae Lim, Alice Oh</p><p><b>Upvotes:</b> 61</p><p><b>Summary:</b> We introduce Solar Open, a 102B-parameter bilingual Mixture-of-Experts language model for underserved languages. Solar Open demonstrates a systematic methodology for building competitive LLMs by addressing three interconnected challenges. First, to train effectively despite data scarcity for underserved languages, we synthesize 4.5T tokens of high-quality, domain-specific, and RL-oriented data. Second, we coordinate this data through a progressive curriculum jointly optimizing composition, quality thresholds, and domain coverage across 20 trillion tokens. Third, to enable reasoning capabilities through scalable RL, we apply our proposed framework SnapPO for efficient optimization. Across benchmarks in English and Korean, Solar Open achieves competitive performance, demonstrating the effectiveness of this methodology for underserved language AI development.</p> - https://arxiv.org/abs/2601.07022 - Sun, 11 Jan 2026 18:33:09 +0000 + sangkuriang: A pseudo-spectral Python library for Korteweg-de Vries soliton simulation + https://arxiv.org/abs/2601.12029 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.12029.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Sandy H. S. Herho, Faruq Khadami, Iwan P. Anwar, Dasapta E. Irawan</p><p><b>Upvotes:</b> 2</p><p><b>Summary:</b> The Korteweg-de Vries (KdV) equation serves as a foundational model in nonlinear wave physics, describing the balance between dispersive spreading and nonlinear steepening that gives rise to solitons. This article introduces sangkuriang, an open-source Python library for solving this equation using Fourier pseudo-spectral spatial discretization coupled with adaptive high-order time integration. The implementation leverages just-in-time (JIT) compilation for computational efficiency while maintaining accessibility for instructional purposes. Validation encompasses progressively complex scenarios including isolated soliton propagation, symmetric two-wave configurations, overtaking collisions between waves of differing amplitudes, and three-body interactions. Conservation of the classical invariants is monitored throughout, with deviations remaining small across all test cases. Measured soliton velocities conform closely to theoretical predictions based on the amplitude-velocity relationship characteristic of integrable systems. Complementary diagnostics drawn from information theory and recurrence analysis confirm that computed solutions preserve the regular phase-space structure expected for completely integrable dynamics. The solver outputs data in standard scientific formats compatible with common analysis tools and generates visualizations of spatiotemporal wave evolution. By combining numerical accuracy with practical accessibility on modest computational resources, sangkuriang offers a platform suitable for both classroom demonstrations of nonlinear wave phenomena and exploratory research into soliton dynamics.</p> + https://arxiv.org/abs/2601.12029 + Sat, 17 Jan 2026 12:12:52 +0000 - Dr. Zero: Self-Evolving Search Agents without Training Data - https://arxiv.org/abs/2601.07055 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07055.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhenrui Yue, Kartikeya Upasani, Xianjun Yang, Suyu Ge, Shaoliang Nie, Yuning Mao, Zhe Liu, Dong Wang</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> As high-quality data becomes increasingly difficult to obtain, data-free self-evolution has emerged as a promising paradigm. This approach allows large language models (LLMs) to autonomously generate and solve complex problems, thereby improving their reasoning capabilities. However, multi-turn search agents struggle in data-free self-evolution due to the limited question diversity and the substantial compute required for multi-step reasoning and tool using. In this work, we introduce Dr. Zero, a framework enabling search agents to effectively self-evolve without any training data. In particular, we design a self-evolution feedback loop where a proposer generates diverse questions to train a solver initialized from the same base model. As the solver evolves, it incentivizes the proposer to produce increasingly difficult yet solvable tasks, thus establishing an automated curriculum to refine both agents. To enhance training efficiency, we also introduce hop-grouped relative policy optimization (HRPO). This method clusters structurally similar questions to construct group-level baselines, effectively minimizing the sampling overhead in evaluating each query's individual difficulty and solvability. Consequently, HRPO significantly reduces the compute requirements for solver training without compromising performance or stability. Extensive experiment results demonstrate that the data-free Dr. Zero matches or surpasses fully supervised search agents, proving that complex reasoning and search capabilities can emerge solely through self-evolution.</p> - https://arxiv.org/abs/2601.07055 - Sun, 11 Jan 2026 20:27:55 +0000 + ToolPRMBench: Evaluating and Advancing Process Reward Models for Tool-using Agents + https://arxiv.org/abs/2601.12294 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.12294.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dawei Li, Yuguang Yao, Zhen Tan, Huan Liu, Ruocheng Guo</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> Reward-guided search methods have demonstrated strong potential in enhancing tool-using agents by effectively guiding sampling and exploration over complex action spaces. As a core design, those search methods utilize process reward models (PRMs) to provide step-level rewards, enabling more fine-grained monitoring. However, there is a lack of systematic and reliable evaluation benchmarks for PRMs in tool-using settings. In this paper, we introduce ToolPRMBench, a large-scale benchmark specifically designed to evaluate PRMs for tool-using agents. ToolPRMBench is built on top of several representative tool-using benchmarks and converts agent trajectories into step-level test cases. Each case contains the interaction history, a correct action, a plausible but incorrect alternative, and relevant tool metadata. We respectively utilize offline sampling to isolate local single-step errors and online sampling to capture realistic multi-step failures from full agent rollouts. A multi-LLM verification pipeline is proposed to reduce label noise and ensure data quality. We conduct extensive experiments across large language models, general PRMs, and tool-specialized PRMs on ToolPRMBench. The results reveal clear differences in PRM effectiveness and highlight the potential of specialized PRMs for tool-using. Code and data will be released at https://github.com/David-Li0406/ToolPRMBench.</p> + https://arxiv.org/abs/2601.12294 + Sun, 18 Jan 2026 07:48:36 +0000 - Lost in the Noise: How Reasoning Models Fail with Contextual Distractors - https://arxiv.org/abs/2601.07226 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07226.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Seongyun Lee, Yongrae Jo, Minju Seo, Moontae Lee, Minjoon Seo</p><p><b>Upvotes:</b> 28</p><p><b>Summary:</b> Recent advances in reasoning models and agentic AI systems have led to an increased reliance on diverse external information. However, this shift introduces input contexts that are inherently noisy, a reality that current sanitized benchmarks fail to capture. We introduce NoisyBench, a comprehensive benchmark that systematically evaluates model robustness across 11 datasets in RAG, reasoning, alignment, and tool-use tasks against diverse noise types, including random documents, irrelevant chat histories, and hard negative distractors. Our evaluation reveals a catastrophic performance drop of up to 80% in state-of-the-art models when faced with contextual distractors. Crucially, we find that agentic workflows often amplify these errors by over-trusting noisy tool outputs, and distractors can trigger emergent misalignment even without adversarial intent. We find that prompting, context engineering, SFT, and outcome-reward only RL fail to ensure robustness; in contrast, our proposed Rationale-Aware Reward (RARE) significantly strengthens resilience by incentivizing the identification of helpful information within noise. Finally, we uncover an inverse scaling trend where increased test-time computation leads to worse performance in noisy settings and demonstrate via attention visualization that models disproportionately focus on distractor tokens, providing vital insights for building the next generation of robust, reasoning-capable agents.</p> - https://arxiv.org/abs/2601.07226 - Mon, 12 Jan 2026 05:43:51 +0000 + MMDeepResearch-Bench: A Benchmark for Multimodal Deep Research Agents + https://arxiv.org/abs/2601.12346 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.12346.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Peizhou Huang, Zixuan Zhong, Zhongwei Wan, Donghao Zhou, Samiul Alam, Xin Wang, Zexin Li, Zhihao Dou, Li Zhu, Jing Xiong, Chaofan Tao, Yan Xu, Dimitrios Dimitriadis, Tuo Zhang, Mi Zhang</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> Deep Research Agents (DRAs) generate citation-rich reports via multi-step search and synthesis, yet existing benchmarks mainly target text-only settings or short-form multimodal QA, missing end-to-end multimodal evidence use. We introduce MMDeepResearch-Bench (MMDR-Bench), a benchmark of 140 expert-crafted tasks across 21 domains, where each task provides an image-text bundle to evaluate multimodal understanding and citation-grounded report generation. Compared to prior setups, MMDR-Bench emphasizes report-style synthesis with explicit evidence use, where models must connect visual artifacts to sourced claims and maintain consistency across narrative, citations, and visual references. We further propose a unified, interpretable evaluation pipeline: Formula-LLM Adaptive Evaluation (FLAE) for report quality, Trustworthy Retrieval-Aligned Citation Evaluation (TRACE) for citation-grounded evidence alignment, and Multimodal Support-Aligned Integrity Check (MOSAIC) for text-visual integrity, each producing fine-grained signals that support error diagnosis beyond a single overall score. Experiments across 25 state-of-the-art models reveal systematic trade-offs between generation quality, citation discipline, and multimodal grounding, highlighting that strong prose alone does not guarantee faithful evidence use and that multimodal integrity remains a key bottleneck for deep research agents.</p> + https://arxiv.org/abs/2601.12346 + Sun, 18 Jan 2026 10:41:33 +0000 - The Confidence Dichotomy: Analyzing and Mitigating Miscalibration in Tool-Use Agents - https://arxiv.org/abs/2601.07264 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07264.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Weihao Xuan, Qingcheng Zeng, Heli Qi, Yunze Xiao, Junjue Wang, Naoto Yokoya</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> Autonomous agents based on large language models (LLMs) are rapidly evolving to handle multi-turn tasks, but ensuring their trustworthiness remains a critical challenge. A fundamental pillar of this trustworthiness is calibration, which refers to an agent's ability to express confidence that reliably reflects its actual performance. While calibration is well-established for static models, its dynamics in tool-integrated agentic workflows remain underexplored. In this work, we systematically investigate verbalized calibration in tool-use agents, revealing a fundamental confidence dichotomy driven by tool type. Specifically, our pilot study identifies that evidence tools (e.g., web search) systematically induce severe overconfidence due to inherent noise in retrieved information, while verification tools (e.g., code interpreters) can ground reasoning through deterministic feedback and mitigate miscalibration. To robustly improve calibration across tool types, we propose a reinforcement learning (RL) fine-tuning framework that jointly optimizes task accuracy and calibration, supported by a holistic benchmark of reward designs. We demonstrate that our trained agents not only achieve superior calibration but also exhibit robust generalization from local training environments to noisy web settings and to distinct domains such as mathematical reasoning. Our results highlight the necessity of domain-specific calibration strategies for tool-use agents. More broadly, this work establishes a foundation for building self-aware agents that can reliably communicate uncertainty in high-stakes, real-world deployments.</p> - https://arxiv.org/abs/2601.07264 - Mon, 12 Jan 2026 07:10:35 +0000 + Agentic Reasoning for Large Language Models + https://arxiv.org/abs/2601.12538 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.12538.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tianxin Wei, Ting-Wei Li, Zhining Liu, Xuying Ning, Ze Yang, Jiaru Zou, Zhichen Zeng, Ruizhong Qiu, Xiao Lin, Dongqi Fu, Zihao Li, Mengting Ai, Duo Zhou, Wenxuan Bao, Yunzhe Li, Gaotang Li, Cheng Qian, Yu Wang, Xiangru Tang, Yin Xiao, Liri Fang, Hui Liu, Xianfeng Tang, Yuji Zhang, Chi Wang, Jiaxuan You, Heng Ji, Hanghang Tong, Jingrui He</p><p><b>Upvotes:</b> 173</p><p><b>Summary:</b> Reasoning is a fundamental cognitive process underlying inference, problem-solving, and decision-making. While large language models (LLMs) demonstrate strong reasoning capabilities in closed-world settings, they struggle in open-ended and dynamic environments. Agentic reasoning marks a paradigm shift by reframing LLMs as autonomous agents that plan, act, and learn through continual interaction. In this survey, we organize agentic reasoning along three complementary dimensions. First, we characterize environmental dynamics through three layers: foundational agentic reasoning, which establishes core single-agent capabilities including planning, tool use, and search in stable environments; self-evolving agentic reasoning, which studies how agents refine these capabilities through feedback, memory, and adaptation; and collective multi-agent reasoning, which extends intelligence to collaborative settings involving coordination, knowledge sharing, and shared goals. Across these layers, we distinguish in-context reasoning, which scales test-time interaction through structured orchestration, from post-training reasoning, which optimizes behaviors via reinforcement learning and supervised fine-tuning. We further review representative agentic reasoning frameworks across real-world applications and benchmarks, including science, robotics, healthcare, autonomous research, and mathematics. This survey synthesizes agentic reasoning methods into a unified roadmap bridging thought and action, and outlines open challenges and future directions, including personalization, long-horizon interaction, world modeling, scalable multi-agent training, and governance for real-world deployment.</p> + https://arxiv.org/abs/2601.12538 + Sun, 18 Jan 2026 18:58:23 +0000 - Controlled Self-Evolution for Algorithmic Code Optimization - https://arxiv.org/abs/2601.07348 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07348.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tu Hu, Ronghao Chen, Shuo Zhang, Jianghao Yin, Mou Xiao Feng, Jingping Liu, Shaolei Zhang, Wenqi Jiang, Yuqi Fang, Sen Hu, Yi Xu, Huacan Wang</p><p><b>Upvotes:</b> 106</p><p><b>Summary:</b> Self-evolution methods enhance code generation through iterative "generate-verify-refine" cycles, yet existing approaches suffer from low exploration efficiency, failing to discover solutions with superior complexity within limited budgets. This inefficiency stems from initialization bias trapping evolution in poor solution regions, uncontrolled stochastic operations lacking feedback guidance, and insufficient experience utilization across tasks. To address these bottlenecks, we propose Controlled Self-Evolution (CSE), which consists of three key components. Diversified Planning Initialization generates structurally distinct algorithmic strategies for broad solution space coverage. Genetic Evolution replaces stochastic operations with feedback-guided mechanisms, enabling targeted mutation and compositional crossover. Hierarchical Evolution Memory captures both successful and failed experiences at inter-task and intra-task levels. Experiments on EffiBench-X demonstrate that CSE consistently outperforms all baselines across various LLM backbones. Furthermore, CSE achieves higher efficiency from early generations and maintains continuous improvement throughout evolution. Our code is publicly available at https://github.com/QuantaAlpha/EvoControl.</p> - https://arxiv.org/abs/2601.07348 - Mon, 12 Jan 2026 09:23:13 +0000 + Being-H0.5: Scaling Human-Centric Robot Learning for Cross-Embodiment Generalization + https://arxiv.org/abs/2601.12993 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.12993.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hao Luo, Ye Wang, Wanpeng Zhang, Sipeng Zheng, Ziheng Xi, Chaoyi Xu, Haiweng Xu, Haoqi Yuan, Chi Zhang, Yiqing Wang, Yicheng Feng, Zongqing Lu</p><p><b>Upvotes:</b> 74</p><p><b>Summary:</b> We introduce Being-H0.5, a foundational Vision-Language-Action (VLA) model designed for robust cross-embodiment generalization across diverse robotic platforms. While existing VLAs often struggle with morphological heterogeneity and data scarcity, we propose a human-centric learning paradigm that treats human interaction traces as a universal "mother tongue" for physical interaction. To support this, we present UniHand-2.0, the largest embodied pre-training recipe to date, comprising over 35,000 hours of multimodal data across 30 distinct robotic embodiments. Our approach introduces a Unified Action Space that maps heterogeneous robot controls into semantically aligned slots, enabling low-resource robots to bootstrap skills from human data and high-resource platforms. Built upon this human-centric foundation, we design a unified sequential modeling and multi-task pre-training paradigm to bridge human demonstrations and robotic execution. Architecturally, Being-H0.5 utilizes a Mixture-of-Transformers design featuring a novel Mixture-of-Flow (MoF) framework to decouple shared motor primitives from specialized embodiment-specific experts. Finally, to make cross-embodiment policies stable in the real world, we introduce Manifold-Preserving Gating for robustness under sensory shift and Universal Async Chunking to universalize chunked control across embodiments with different latency and control profiles. We empirically demonstrate that Being-H0.5 achieves state-of-the-art results on simulated benchmarks, such as LIBERO (98.9%) and RoboCasa (53.9%), while also exhibiting strong cross-embodiment capabilities on five robotic platforms.</p> + https://arxiv.org/abs/2601.12993 + Mon, 19 Jan 2026 12:20:38 +0000 - Beyond Hard Masks: Progressive Token Evolution for Diffusion Language Models - https://arxiv.org/abs/2601.07351 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07351.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Linhao Zhong, Linyu Wu, Bozhen Fang, Tianjian Feng, Chenchen Jing, Wen Wang, Jiaheng Zhang, Hao Chen, Chunhua Shen</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> Diffusion Language Models (DLMs) offer a promising alternative for language modeling by enabling parallel decoding through iterative refinement. However, most DLMs rely on hard binary masking and discrete token assignments, which hinder the revision of early decisions and underutilize intermediate probabilistic representations. In this paper, we propose EvoToken-DLM, a novel diffusion-based language modeling approach that replaces hard binary masks with evolving soft token distributions. EvoToken-DLM enables a progressive transition from masked states to discrete outputs, supporting revisable decoding. To effectively support this evolution, we introduce continuous trajectory supervision, which aligns training objectives with iterative probabilistic updates. Extensive experiments across multiple benchmarks show that EvoToken-DLM consistently achieves superior performance, outperforming strong diffusion-based and masked DLM baselines. Project webpage: https://aim-uofa.github.io/EvoTokenDLM.</p> - https://arxiv.org/abs/2601.07351 - Mon, 12 Jan 2026 09:25:14 +0000 + Think3D: Thinking with Space for Spatial Reasoning + https://arxiv.org/abs/2601.13029 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13029.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zaibin Zhang, Yuhan Wu, Lianjie Jia, Yifan Wang, Zhongbo Zhang, Yijiang Li, Binghao Ran, Fuxi Zhang, Zhuohan Sun, Zhenfei Yin, Lijun Wang, Huchuan Lu</p><p><b>Upvotes:</b> 45</p><p><b>Summary:</b> Understanding and reasoning about the physical world requires spatial intelligence: the ability to interpret geometry, perspective, and spatial relations beyond 2D perception. While recent vision large models (VLMs) excel at visual understanding, they remain fundamentally 2D perceivers and struggle with genuine 3D reasoning. We introduce Think3D, a framework that enables VLM agents to think with 3D space. By leveraging 3D reconstruction models that recover point clouds and camera poses from images or videos, Think3D allows the agent to actively manipulate space through camera-based operations and ego/global-view switching, transforming spatial reasoning into an interactive 3D chain-of-thought process. Without additional training, Think3D significantly improves the spatial reasoning performance of advanced models such as GPT-4.1 and Gemini 2.5 Pro, yielding average gains of +7.8% on BLINK Multi-view and MindCube, and +4.7% on VSI-Bench. We further show that smaller models, which struggle with spatial exploration, benefit significantly from a reinforcement learning policy that enables the model to select informative viewpoints and operations. With RL, the benefit from tool usage increases from +0.7% to +6.8%. Our findings demonstrate that training-free, tool-augmented spatial exploration is a viable path toward more flexible and human-like 3D reasoning in multimodal agents, establishing a new dimension of multimodal intelligence. Code and weights are released at https://github.com/zhangzaibin/spagent.</p> + https://arxiv.org/abs/2601.13029 + Mon, 19 Jan 2026 13:13:54 +0000 - MegaFlow: Large-Scale Distributed Orchestration System for the Agentic Era - https://arxiv.org/abs/2601.07526 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07526.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lei Zhang, Mouxiang Chen, Ruisheng Cao, Jiawei Chen, Fan Zhou, Yiheng Xu, Jiaxi Yang, Liang Chen, Changwei Luo, Kai Zhang, Fan Yan, KaShun Shum, Jiajun Zhang, Zeyu Cui, Hu Feng, Junyang Lin, Binyuan Hui, Min Yang</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> The rapid development of interactive and autonomous AI systems signals our entry into the agentic era. Training and evaluating agents on complex agentic tasks such as software engineering and computer use requires not only efficient model computation but also sophisticated infrastructure capable of coordinating vast agent-environment interactions. However, no open-source infrastructure can effectively support large-scale training and evaluation on such complex agentic tasks. To address this challenge, we present MegaFlow, a large-scale distributed orchestration system that enables efficient scheduling, resource allocation, and fine-grained task management for agent-environment workloads. MegaFlow abstracts agent training infrastructure into three independent services (Model Service, Agent Service, and Environment Service) that interact through unified interfaces, enabling independent scaling and flexible resource allocation across diverse agent-environment configurations. In our agent training deployments, MegaFlow successfully orchestrates tens of thousands of concurrent agent tasks while maintaining high system stability and achieving efficient resource utilization. By enabling such large-scale agent training, MegaFlow addresses a critical infrastructure gap in the emerging agentic AI landscape.</p> - https://arxiv.org/abs/2601.07526 - Mon, 12 Jan 2026 13:25:33 +0000 + Typhoon ASR Real-time: FastConformer-Transducer for Thai Automatic Speech Recognition + https://arxiv.org/abs/2601.13044 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13044.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Warit Sirichotedumrong, Adisai Na-Thalang, Potsawee Manakul, Pittawat Taveekitworachai, Sittipong Sripaisarnmongkol, Kunat Pipatanakul</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Large encoder-decoder models like Whisper achieve strong offline transcription but remain impractical for streaming applications due to high latency. However, due to the accessibility of pre-trained checkpoints, the open Thai ASR landscape remains dominated by these offline architectures, leaving a critical gap in efficient streaming solutions. We present Typhoon ASR Real-time, a 115M-parameter FastConformer-Transducer model for low-latency Thai speech recognition. We demonstrate that rigorous text normalization can match the impact of model scaling: our compact model achieves a 45x reduction in computational cost compared to Whisper Large-v3 while delivering comparable accuracy. Our normalization pipeline resolves systemic ambiguities in Thai transcription --including context-dependent number verbalization and repetition markers (mai yamok) --creating consistent training targets. We further introduce a two-stage curriculum learning approach for Isan (north-eastern) dialect adaptation that preserves Central Thai performance. To address reproducibility challenges in Thai ASR, we release the Typhoon ASR Benchmark, a gold-standard human-labeled datasets with transcriptions following established Thai linguistic conventions, providing standardized evaluation protocols for the research community.</p> + https://arxiv.org/abs/2601.13044 + Mon, 19 Jan 2026 13:28:17 +0000 - TAG-MoE: Task-Aware Gating for Unified Generative Mixture-of-Experts - https://arxiv.org/abs/2601.08881 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08881.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yu Xu, Hongbin Yan, Juan Cao, Yiji Cheng, Tiankai Hang, Runze He, Zijin Yin, Shiyi Zhang, Yuxin Zhang, Jintao Li, Chunyu Wang, Qinglin Lu, Tong-Yee Lee, Fan Tang</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Unified image generation and editing models suffer from severe task interference in dense diffusion transformers architectures, where a shared parameter space must compromise between conflicting objectives (e.g., local editing v.s. subject-driven generation). While the sparse Mixture-of-Experts (MoE) paradigm is a promising solution, its gating networks remain task-agnostic, operating based on local features, unaware of global task intent. This task-agnostic nature prevents meaningful specialization and fails to resolve the underlying task interference. In this paper, we propose a novel framework to inject semantic intent into MoE routing. We introduce a Hierarchical Task Semantic Annotation scheme to create structured task descriptors (e.g., scope, type, preservation). We then design Predictive Alignment Regularization to align internal routing decisions with the task's high-level semantics. This regularization evolves the gating network from a task-agnostic executor to a dispatch center. Our model effectively mitigates task interference, outperforming dense baselines in fidelity and quality, and our analysis shows that experts naturally develop clear and semantically correlated specializations.</p> - https://arxiv.org/abs/2601.08881 - Mon, 12 Jan 2026 14:52:45 +0000 + Aligning Agentic World Models via Knowledgeable Experience Learning + https://arxiv.org/abs/2601.13247 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13247.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Baochang Ren, Yunzhi Yao, Rui Sun, Shuofei Qiao, Ningyu Zhang, Huajun Chen</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Current Large Language Models (LLMs) exhibit a critical modal disconnect: they possess vast semantic knowledge but lack the procedural grounding to respect the immutable laws of the physical world. Consequently, while these agents implicitly function as world models, their simulations often suffer from physical hallucinations-generating plans that are logically sound but physically unexecutable. Existing alignment strategies predominantly rely on resource-intensive training or fine-tuning, which attempt to compress dynamic environmental rules into static model parameters. However, such parametric encapsulation is inherently rigid, struggling to adapt to the open-ended variability of physical dynamics without continuous, costly retraining. To bridge this gap, we introduce WorldMind, a framework that autonomously constructs a symbolic World Knowledge Repository by synthesizing environmental feedback. Specifically, it unifies Process Experience to enforce physical feasibility via prediction errors and Goal Experience to guide task optimality through successful trajectories. Experiments on EB-ALFRED and EB-Habitat demonstrate that WorldMind achieves superior performance compared to baselines with remarkable cross-model and cross-environment transferability.</p> + https://arxiv.org/abs/2601.13247 + Mon, 19 Jan 2026 17:33:31 +0000 - Beyond Static Tools: Test-Time Tool Evolution for Scientific Reasoning - https://arxiv.org/abs/2601.07641 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07641.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaxuan Lu, Ziyu Kong, Yemin Wang, Rong Fu, Haiyuan Wan, Cheng Yang, Wenjie Lou, Haoran Sun, Lilong Wang, Yankai Jiang, Xiaosong Wang, Xiao Sun, Dongzhan Zhou</p><p><b>Upvotes:</b> 42</p><p><b>Summary:</b> The central challenge of AI for Science is not reasoning alone, but the ability to create computational methods in an open-ended scientific world. Existing LLM-based agents rely on static, pre-defined tool libraries, a paradigm that fundamentally fails in scientific domains where tools are sparse, heterogeneous, and intrinsically incomplete. In this paper, we propose Test-Time Tool Evolution (TTE), a new paradigm that enables agents to synthesize, verify, and evolve executable tools during inference. By transforming tools from fixed resources into problem-driven artifacts, TTE overcomes the rigidity and long-tail limitations of static tool libraries. To facilitate rigorous evaluation, we introduce SciEvo, a benchmark comprising 1,590 scientific reasoning tasks supported by 925 automatically evolved tools. Extensive experiments show that TTE achieves state-of-the-art performance in both accuracy and tool efficiency, while enabling effective cross-domain adaptation of computational tools. The code and benchmark have been released at https://github.com/lujiaxuan0520/Test-Time-Tool-Evol.</p> - https://arxiv.org/abs/2601.07641 - Mon, 12 Jan 2026 15:22:51 +0000 + Beyond Cosine Similarity: Taming Semantic Drift and Antonym Intrusion in a 15-Million Node Turkish Synonym Graph + https://arxiv.org/abs/2601.13251 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13251.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ebubekir Tosun, Mehmet Emin Buldur, Özay Ezerceli, Mahmoud ElHussieni</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> Neural embeddings have a notorious blind spot: they can't reliably tell synonyms apart from antonyms. Consequently, increasing similarity thresholds often fails to prevent opposites from being grouped together. We've built a large-scale semantic clustering system specifically designed to tackle this problem head on. Our pipeline chews through 15 million lexical items, evaluates a massive 520 million potential relationships, and ultimately generates 2.9 million high-precision semantic clusters. The system makes three primary contributions. First, we introduce a labeled dataset of 843,000 concept pairs spanning synonymy, antonymy, and co-hyponymy, constructed via Gemini 2.5-Flash LLM augmentation and verified using human-curated dictionary resources. Second, we propose a specialized three-way semantic relation discriminator that achieves 90% macro-F1, enabling robust disambiguation beyond raw embedding similarity. Third, we introduce a novel soft-to-hard clustering algorithm that mitigates semantic drift preventing erroneous transitive chains (e.g., hot -> spicy -> pain -> depression) while simultaneously resolving polysemy. Our approach employs a topology-aware two-stage expansion-pruning procedure with topological voting, ensuring that each term is assigned to exactly one semantically coherent cluster. The resulting resource enables high-precision semantic search and retrieval-augmented generation, particularly for morphologically rich and low-resource languages where existing synonym databases remain sparse.</p> + https://arxiv.org/abs/2601.13251 + Mon, 19 Jan 2026 17:37:25 +0000 - OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent - https://arxiv.org/abs/2601.07779 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07779.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Bowen Yang, Kaiming Jin, Zhenyu Wu, Zhaoyang Liu, Qiushi Sun, Zehao Li, JingJing Xie, Zhoumianze Liu, Fangzhi Xu, Kanzhi Cheng, Qingyun Li, Yian Wang, Yu Qiao, Zun Wang, Zichen Ding</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> While Vision-Language Models (VLMs) have significantly advanced Computer-Using Agents (CUAs), current frameworks struggle with robustness in long-horizon workflows and generalization in novel domains. These limitations stem from a lack of granular control over historical visual context curation and the absence of visual-aware tutorial retrieval. To bridge these gaps, we introduce OS-Symphony, a holistic framework that comprises an Orchestrator coordinating two key innovations for robust automation: (1) a Reflection-Memory Agent that utilizes milestone-driven long-term memory to enable trajectory-level self-correction, effectively mitigating visual context loss in long-horizon tasks; (2) Versatile Tool Agents featuring a Multimodal Searcher that adopts a SeeAct paradigm to navigate a browser-based sandbox to synthesize live, visually aligned tutorials, thereby resolving fidelity issues in unseen scenarios. Experimental results demonstrate that OS-Symphony delivers substantial performance gains across varying model scales, establishing new state-of-the-art results on three online benchmarks, notably achieving 65.84% on OSWorld.</p> - https://arxiv.org/abs/2601.07779 - Mon, 12 Jan 2026 17:55:51 +0000 + A Hybrid Protocol for Large-Scale Semantic Dataset Generation in Low-Resource Languages: The Turkish Semantic Relations Corpus + https://arxiv.org/abs/2601.13253 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13253.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ebubekir Tosun, Mehmet Emin Buldur, Özay Ezerceli, Mahmoud ElHussieni</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> We present a hybrid methodology for generating large-scale semantic relationship datasets in low-resource languages, demonstrated through a comprehensive Turkish semantic relations corpus. Our approach integrates three phases: (1) FastText embeddings with Agglomerative Clustering to identify semantic clusters, (2) Gemini 2.5-Flash for automated semantic relationship classification, and (3) integration with curated dictionary sources. The resulting dataset comprises 843,000 unique Turkish semantic pairs across three relationship types (synonyms, antonyms, co-hyponyms) representing a 10x scale increase over existing resources at minimal cost ($65). We validate the dataset through two downstream tasks: an embedding model achieving 90% top-1 retrieval accuracy and a classification model attaining 90% F1-macro. Our scalable protocol addresses critical data scarcity in Turkish NLP and demonstrates applicability to other low-resource languages. We publicly release the dataset and models.</p> + https://arxiv.org/abs/2601.13253 + Mon, 19 Jan 2026 17:38:52 +0000 - MHLA: Restoring Expressivity of Linear Attention via Token-Level Multi-Head - https://arxiv.org/abs/2601.07832 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.07832.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Kewei Zhang, Ye Huang, Yufan Deng, Jincheng Yu, Junsong Chen, Huan Ling, Enze Xie, Daquan Zhou</p><p><b>Upvotes:</b> 44</p><p><b>Summary:</b> While the Transformer architecture dominates many fields, its quadratic self-attention complexity hinders its use in large-scale applications. Linear attention offers an efficient alternative, but its direct application often degrades performance, with existing fixes typically re-introducing computational overhead through extra modules (e.g., depthwise separable convolution) that defeat the original purpose. In this work, we identify a key failure mode in these methods: global context collapse, where the model loses representational diversity. To address this, we propose Multi-Head Linear Attention (MHLA), which preserves this diversity by computing attention within divided heads along the token dimension. We prove that MHLA maintains linear complexity while recovering much of the expressive power of softmax attention, and verify its effectiveness across multiple domains, achieving a 3.6\% improvement on ImageNet classification, a 6.3\% gain on NLP, a 12.6\% improvement on image generation, and a 41\% enhancement on video generation under the same time complexity.</p> - https://arxiv.org/abs/2601.07832 - Mon, 12 Jan 2026 18:59:18 +0000 + A BERTology View of LLM Orchestrations: Token- and Layer-Selective Probes for Efficient Single-Pass Classification + https://arxiv.org/abs/2601.13288 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13288.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Gonzalo Ariel Meyoyan, Luciano Del Corro</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Production LLM systems often rely on separate models for safety and other classification-heavy steps, increasing latency, VRAM footprint, and operational complexity. We instead reuse computation already paid for by the serving LLM: we train lightweight probes on its hidden states and predict labels in the same forward pass used for generation. We frame classification as representation selection over the full token-layer hidden-state tensor, rather than committing to a fixed token or fixed layer (e.g., first-token logits or final-layer pooling). To implement this, we introduce a two-stage aggregator that (i) summarizes tokens within each layer and (ii) aggregates across layer summaries to form a single representation for classification. We instantiate this template with direct pooling, a 100K-parameter scoring-attention gate, and a downcast multi-head self-attention (MHA) probe with up to 35M trainable parameters. Across safety and sentiment benchmarks our probes improve over logit-only reuse (e.g., MULI) and are competitive with substantially larger task-specific baselines, while preserving near-serving latency and avoiding the VRAM and latency costs of a separate guard-model pipeline.</p> + https://arxiv.org/abs/2601.13288 + Mon, 19 Jan 2026 18:40:29 +0000 - MemoBrain: Executive Memory as an Agentic Brain for Reasoning - https://arxiv.org/abs/2601.08079 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08079.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hongjin Qian, Zhao Cao, Zheng Liu</p><p><b>Upvotes:</b> 35</p><p><b>Summary:</b> Complex reasoning in tool-augmented agent frameworks is inherently long-horizon, causing reasoning traces and transient tool artifacts to accumulate and strain the bounded working context of large language models. Without explicit memory mechanisms, such accumulation disrupts logical continuity and undermines task alignment. This positions memory not as an auxiliary efficiency concern, but as a core component for sustaining coherent, goal-directed reasoning over long horizons. We propose MemoBrain, an executive memory model for tool-augmented agents that constructs a dependency-aware memory over reasoning steps, capturing salient intermediate states and their logical relations. Operating as a co-pilot alongside the reasoning agent, MemoBrain organizes reasoning progress without blocking execution and actively manages the working context. Specifically, it prunes invalid steps, folds completed sub-trajectories, and preserves a compact, high-salience reasoning backbone under a fixed context budget. Together, these mechanisms enable explicit cognitive control over reasoning trajectories rather than passive context accumulation. We evaluate MemoBrain on challenging long-horizon benchmarks, including GAIA, WebWalker, and BrowseComp-Plus, demonstrating consistent improvements over strong baselines.</p> - https://arxiv.org/abs/2601.08079 - Mon, 12 Jan 2026 23:44:59 +0000 + Behavior Knowledge Merge in Reinforced Agentic Models + https://arxiv.org/abs/2601.13572 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13572.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiangchi Yuan, Dachuan Shi, Chunhui Zhang, Zheyuan Liu, Shenglong Yao, Soroush Vosoughi, Wenke Lee</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> Reinforcement learning (RL) is central to post-training, particularly for agentic models that require specialized reasoning behaviors. In this setting, model merging offers a practical mechanism for integrating multiple RL-trained agents from different tasks into a single generalist model. However, existing merging methods are designed for supervised fine-tuning (SFT), and they are suboptimal to preserve task-specific capabilities on RL-trained agentic models. The root is a task-vector mismatch between RL and SFT: on-policy RL induces task vectors that are highly sparse and heterogeneous, whereas SFT-style merging implicitly assumes dense and globally comparable task vectors. When standard global averaging is applied under this mismatch, RL's non-overlapping task vectors that encode critical task-specific behaviors are reduced and parameter updates are diluted. To address this issue, we propose Reinforced Agent Merging (RAM), a distribution-aware merging framework explicitly designed for RL-trained agentic models. RAM disentangles shared and task-specific unique parameter updates, averaging shared components while selectively preserving and rescaling unique ones to counteract parameter update dilution. Experiments across multiple agent domains and model architectures demonstrate that RAM not only surpasses merging baselines, but also unlocks synergistic potential among agents to achieve performance superior to that of specialized agents in their domains.</p> + https://arxiv.org/abs/2601.13572 + Tue, 20 Jan 2026 03:56:53 +0000 - The Agent's First Day: Benchmarking Learning, Exploration, and Scheduling in the Workplace Scenarios - https://arxiv.org/abs/2601.08173 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08173.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Daocheng Fu, Jianbiao Mei, Rong Wu, Xuemeng Yang, Jia Xu, Ding Wang, Pinlong Cai, Yong Liu, Licheng Wen, Botian Shi</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> The rapid evolution of Multi-modal Large Language Models (MLLMs) has advanced workflow automation; however, existing research mainly targets performance upper bounds in static environments, overlooking robustness for stochastic real-world deployment. We identify three key challenges: dynamic task scheduling, active exploration under uncertainty, and continuous learning from experience. To bridge this gap, we introduce , a dynamic evaluation environment that simulates a "trainee" agent continuously exploring a novel setting. Unlike traditional benchmarks, evaluates agents along three dimensions: (1) context-aware scheduling for streaming tasks with varying priorities; (2) prudent information acquisition to reduce hallucination via active exploration; and (3) continuous evolution by distilling generalized strategies from rule-based, dynamically generated tasks. Experiments show that cutting-edge agents have significant deficiencies in dynamic environments, especially in active exploration and continual learning. Our work establishes a framework for assessing agent reliability, shifting evaluation from static tests to realistic, production-oriented scenarios. Our codes are available at https://github.com/KnowledgeXLab/EvoEnv</p> - https://arxiv.org/abs/2601.08173 - Tue, 13 Jan 2026 03:09:18 +0000 + Uncertainty-Aware Gradient Signal-to-Noise Data Selection for Instruction Tuning + https://arxiv.org/abs/2601.13697 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13697.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhihang Yuan, Chengyu Yue, Long Huang, Litu Ou, Lei Shi</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> Instruction tuning is a standard paradigm for adapting large language models (LLMs), but modern instruction datasets are large, noisy, and redundant, making full-data fine-tuning costly and often unnecessary. Existing data selection methods either build expensive gradient datastores or assign static scores from a weak proxy, largely ignoring evolving uncertainty, and thus missing a key source of LLM interpretability. We propose GRADFILTERING, an objective-agnostic, uncertainty-aware data selection framework that utilizes a small GPT-2 proxy with a LoRA ensemble and aggregates per-example gradients into a Gradient Signal-to-Noise Ratio (G-SNR) utility. Our method matches or surpasses random subsets and strong baselines in most LLM-as-a-judge evaluations as well as in human assessment. Moreover, GRADFILTERING-selected subsets converge faster than competitive filters under the same compute budget, reflecting the benefit of uncertainty-aware scoring.</p> + https://arxiv.org/abs/2601.13697 + Tue, 20 Jan 2026 07:51:32 +0000 - User-Oriented Multi-Turn Dialogue Generation with Tool Use at scale - https://arxiv.org/abs/2601.08225 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08225.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jungho Cho, Minbyul Jeong, Sungrae Park</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> The recent paradigm shift toward large reasoning models (LRMs) as autonomous agents has intensified the demand for sophisticated, multi-turn tool-use capabilities. Yet, existing datasets and data-generation approaches are limited by static, predefined toolsets that cannot scale to the complexity of open-ended human-agent collaboration. To address this, we initially developed a framework for automated task-oriented multi-turn dialogue generation at scale, utilizing an LRM-based simulator to dynamically generate high-value, domain-specific tools to solve specified tasks. However, we observe that a purely task-oriented design often results in "solely task-solving" trajectories, where the agent completes the objective with minimal interaction, failing to generate the high turn-count conversations seen in realistic scenarios. To bridge this gap, we shift toward a user-oriented simulation paradigm. By decoupling task generation from a dedicated user simulator that mimics human behavioral rules - such as incremental request-making and turn-by-turn feedback - we facilitate more authentic, extended multi-turn dialogues that reflect the iterative nature of real-world problem solving. Our generation pipeline operates as a versatile, plug-and-play module capable of initiating generation from any state, ensuring high scalability in producing extended tool-use data. Furthermore, by facilitating multiple task completions within a single trajectory, it yields a high-density dataset that reflects the multifaceted demands of real-world human-agent interaction.</p> - https://arxiv.org/abs/2601.08225 - Tue, 13 Jan 2026 05:14:09 +0000 + DARC: Decoupled Asymmetric Reasoning Curriculum for LLM Evolution + https://arxiv.org/abs/2601.13761 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13761.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shengda Fan, Xuyan Ye, Yankai Lin</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Self-play with large language models has emerged as a promising paradigm for achieving self-improving artificial intelligence. However, existing self-play frameworks often suffer from optimization instability, due to (i) non-stationary objectives induced by solver-dependent reward feedback for the Questioner, and (ii) bootstrapping errors from self-generated pseudo-labels used to supervise the Solver. To mitigate these challenges, we introduce DARC (Decoupled Asymmetric Reasoning Curriculum), a two-stage framework that stabilizes the self-evolution process. First, we train the Questioner to synthesize difficulty-calibrated questions, conditioned on explicit difficulty levels and external corpora. Second, we train the Solver with an asymmetric self-distillation mechanism, where a document-augmented teacher generates high-quality pseudo-labels to supervise the student Solver that lacks document access. Empirical results demonstrate that DARC is model-agnostic, yielding an average improvement of 10.9 points across nine reasoning benchmarks and three backbone models. Moreover, DARC consistently outperforms all baselines and approaches the performance of fully supervised models without relying on human annotations.The code is available at https://github.com/RUCBM/DARC.</p> + https://arxiv.org/abs/2601.13761 + Tue, 20 Jan 2026 09:12:27 +0000 - SnapGen++: Unleashing Diffusion Transformers for Efficient High-Fidelity Image Generation on Edge Devices - https://arxiv.org/abs/2601.08303 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08303.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dongting Hu, Aarush Gupta, Magzhan Gabidolla, Arpit Sahni, Huseyin Coskun, Yanyu Li, Yerlan Idelbayev, Ahsan Mahmood, Aleksei Lebedev, Dishani Lahiri, Anujraaj Goyal, Ju Hu, Mingming Gong, Sergey Tulyakov, Anil Kag</p><p><b>Upvotes:</b> 13</p><p><b>Summary:</b> Recent advances in diffusion transformers (DiTs) have set new standards in image generation, yet remain impractical for on-device deployment due to their high computational and memory costs. In this work, we present an efficient DiT framework tailored for mobile and edge devices that achieves transformer-level generation quality under strict resource constraints. Our design combines three key components. First, we propose a compact DiT architecture with an adaptive global-local sparse attention mechanism that balances global context modeling and local detail preservation. Second, we propose an elastic training framework that jointly optimizes sub-DiTs of varying capacities within a unified supernetwork, allowing a single model to dynamically adjust for efficient inference across different hardware. Finally, we develop Knowledge-Guided Distribution Matching Distillation, a step-distillation pipeline that integrates the DMD objective with knowledge transfer from few-step teacher models, producing high-fidelity and low-latency generation (e.g., 4-step) suitable for real-time on-device use. Together, these contributions enable scalable, efficient, and high-quality diffusion models for deployment on diverse hardware.</p> - https://arxiv.org/abs/2601.08303 - Tue, 13 Jan 2026 07:46:46 +0000 + FutureOmni: Evaluating Future Forecasting from Omni-Modal Context for Multimodal LLMs + https://arxiv.org/abs/2601.13836 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13836.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qian Chen, Jinlan Fu, Changsong Li, See-Kiong Ng, Xipeng Qiu</p><p><b>Upvotes:</b> 34</p><p><b>Summary:</b> Although Multimodal Large Language Models (MLLMs) demonstrate strong omni-modal perception, their ability to forecast future events from audio-visual cues remains largely unexplored, as existing benchmarks focus mainly on retrospective understanding. To bridge this gap, we introduce FutureOmni, the first benchmark designed to evaluate omni-modal future forecasting from audio-visual environments. The evaluated models are required to perform cross-modal causal and temporal reasoning, as well as effectively leverage internal knowledge to predict future events. FutureOmni is constructed via a scalable LLM-assisted, human-in-the-loop pipeline and contains 919 videos and 1,034 multiple-choice QA pairs across 8 primary domains. Evaluations on 13 omni-modal and 7 video-only models show that current systems struggle with audio-visual future prediction, particularly in speech-heavy scenarios, with the best accuracy of 64.8% achieved by Gemini 3 Flash. To mitigate this limitation, we curate a 7K-sample instruction-tuning dataset and propose an Omni-Modal Future Forecasting (OFF) training strategy. Evaluations on FutureOmni and popular audio-visual and video-only benchmarks demonstrate that OFF enhances future forecasting and generalization. We publicly release all code (https://github.com/OpenMOSS/FutureOmni) and datasets (https://huggingface.co/datasets/OpenMOSS-Team/FutureOmni).</p> + https://arxiv.org/abs/2601.13836 + Tue, 20 Jan 2026 10:47:20 +0000 - UM-Text: A Unified Multimodal Model for Image Understanding - https://arxiv.org/abs/2601.08321 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08321.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lichen Ma, Xiaolong Fu, Gaojing Zhou, Zipeng Guo, Ting Zhu, Yichun Liu, Yu Shi, Jason Li, Junshi Huang</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> With the rapid advancement of image generation, visual text editing using natural language instructions has received increasing attention. The main challenge of this task is to fully understand the instruction and reference image, and thus generate visual text that is style-consistent with the image. Previous methods often involve complex steps of specifying the text content and attributes, such as font size, color, and layout, without considering the stylistic consistency with the reference image. To address this, we propose UM-Text, a unified multimodal model for context understanding and visual text editing by natural language instructions. Specifically, we introduce a Visual Language Model (VLM) to process the instruction and reference image, so that the text content and layout can be elaborately designed according to the context information. To generate an accurate and harmonious visual text image, we further propose the UM-Encoder to combine the embeddings of various condition information, where the combination is automatically configured by VLM according to the input instruction. During training, we propose a regional consistency loss to offer more effective supervision for glyph generation on both latent and RGB space, and design a tailored three-stage training strategy to further enhance model performance. In addition, we contribute the UM-DATA-200K, a large-scale visual text image dataset on diverse scenes for model training. Extensive qualitative and quantitative results on multiple public benchmarks demonstrate that our method achieves state-of-the-art performance.</p> - https://arxiv.org/abs/2601.08321 - Tue, 13 Jan 2026 08:18:49 +0000 + AgentEHR: Advancing Autonomous Clinical Decision-Making via Retrospective Summarization + https://arxiv.org/abs/2601.13918 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13918.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yusheng Liao, Chuan Xuan, Yutong Cai, Lina Yang, Zhe Chen, Yanfeng Wang, Yu Wang</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Large Language Models have demonstrated profound utility in the medical domain. However, their application to autonomous Electronic Health Records~(EHRs) navigation remains constrained by a reliance on curated inputs and simplified retrieval tasks. To bridge the gap between idealized experimental settings and realistic clinical environments, we present AgentEHR. This benchmark challenges agents to execute complex decision-making tasks, such as diagnosis and treatment planning, requiring long-range interactive reasoning directly within raw and high-noise databases. In tackling these tasks, we identify that existing summarization methods inevitably suffer from critical information loss and fractured reasoning continuity. To address this, we propose RetroSum, a novel framework that unifies a retrospective summarization mechanism with an evolving experience strategy. By dynamically re-evaluating interaction history, the retrospective mechanism prevents long-context information loss and ensures unbroken logical coherence. Additionally, the evolving strategy bridges the domain gap by retrieving accumulated experience from a memory bank. Extensive empirical evaluations demonstrate that RetroSum achieves performance gains of up to 29.16% over competitive baselines, while significantly decreasing total interaction errors by up to 92.3%.</p> + https://arxiv.org/abs/2601.13918 + Tue, 20 Jan 2026 12:48:04 +0000 - JudgeRLVR: Judge First, Generate Second for Efficient Reasoning - https://arxiv.org/abs/2601.08468 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08468.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiangshan Duo, Hanyu Li, Hailin Zhang, Yudong Wang, Sujian Li, Liang Zhao</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Reinforcement Learning with Verifiable Rewards (RLVR) has become a standard paradigm for reasoning in Large Language Models. However, optimizing solely for final-answer correctness often drives models into aimless, verbose exploration, where they rely on exhaustive trial-and-error tactics rather than structured planning to reach solutions. While heuristic constraints like length penalties can reduce verbosity, they often truncate essential reasoning steps, creating a difficult trade-off between efficiency and verification. In this paper, we argue that discriminative capability is a prerequisite for efficient generation: by learning to distinguish valid solutions, a model can internalize a guidance signal that prunes the search space. We propose JudgeRLVR, a two-stage judge-then-generate paradigm. In the first stage, we train the model to judge solution responses with verifiable answers. In the second stage, we fine-tune the same model with vanilla generating RLVR initialized from the judge. Compared to Vanilla RLVR using the same math-domain training data, JudgeRLVR achieves a better quality--efficiency trade-off for Qwen3-30B-A3B: on in-domain math, it delivers about +3.7 points average accuracy gain with -42\% average generation length; on out-of-domain benchmarks, it delivers about +4.5 points average accuracy improvement, demonstrating enhanced generalization.</p> - https://arxiv.org/abs/2601.08468 - Tue, 13 Jan 2026 11:47:42 +0000 + FantasyVLN: Unified Multimodal Chain-of-Thought Reasoning for Vision-Language Navigation + https://arxiv.org/abs/2601.13976 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.13976.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jing Zuo, Lingzhou Mu, Fan Jiang, Chengcheng Ma, Mu Xu, Yonggang Qi</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> Achieving human-level performance in Vision-and-Language Navigation (VLN) requires an embodied agent to jointly understand multimodal instructions and visual-spatial context while reasoning over long action sequences. Recent works, such as NavCoT and NavGPT-2, demonstrate the potential of Chain-of-Thought (CoT) reasoning for improving interpretability and long-horizon planning. Moreover, multimodal extensions like OctoNav-R1 and CoT-VLA further validate CoT as a promising pathway toward human-like navigation reasoning. However, existing approaches face critical drawbacks: purely textual CoTs lack spatial grounding and easily overfit to sparse annotated reasoning steps, while multimodal CoTs incur severe token inflation by generating imagined visual observations, making real-time navigation impractical. In this work, we propose FantasyVLN, a unified implicit reasoning framework that preserves the benefits of CoT reasoning without explicit token overhead. Specifically, imagined visual tokens are encoded into a compact latent space using a pretrained Visual AutoRegressor (VAR) during CoT reasoning training, and the model jointly learns from textual, visual, and multimodal CoT modes under a unified multi-CoT strategy. At inference, our model performs direct instruction-to-action mapping while still enjoying reasoning-aware representations. Extensive experiments on LH-VLN show that our approach achieves reasoning-aware yet real-time navigation, improving success rates and efficiency while reducing inference latency by an order of magnitude compared to explicit CoT methods.</p> + https://arxiv.org/abs/2601.13976 + Tue, 20 Jan 2026 13:54:10 +0000 - Ministral 3 - https://arxiv.org/abs/2601.08584 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08584.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Alexander H. Liu, Kartik Khandelwal, Sandeep Subramanian, Victor Jouault, Abhinav Rastogi, Adrien Sadé, Alan Jeffares, Albert Jiang, Alexandre Cahill, Alexandre Gavaudan, Alexandre Sablayrolles, Amélie Héliou, Amos You, Andy Ehrenberg, Andy Lo, Anton Eliseev, Antonia Calvi, Avinash Sooriyarachchi, Baptiste Bout, Baptiste Rozière, Baudouin De Monicault, Clémence Lanfranchi, Corentin Barreau, Cyprien Courtot, Daniele Grattarola, Darius Dabert, Diego de las Casas, Elliot Chane-Sane, Faruk Ahmed, Gabrielle Berrada, Gaëtan Ecrepont, Gauthier Guinet, Georgii Novikov, Guillaume Kunsch, Guillaume Lample, Guillaume Martin, Gunshi Gupta, Jan Ludziejewski, Jason Rute, Joachim Studnia, Jonas Amar, Joséphine Delas, Josselin Somerville Roberts, Karmesh Yadav, Khyathi Chandu, Kush Jain, Laurence Aitchison, Laurent Fainsin, Léonard Blier, Lingxiao Zhao, Louis Martin, Lucile Saulnier, Luyu Gao, Maarten Buyl, Margaret Jennings, Marie Pellat, Mark Prins, Mathieu Poirée, Mathilde Guillaumin, Matthieu Dinot, Matthieu Futeral, Maxime Darrin, Maximilian Augustin, Mia Chiquier, Michel Schimpf, Nathan Grinsztajn, Neha Gupta, Nikhil Raghuraman, Olivier Bousquet, Olivier Duchenne, Patricia Wang, Patrick von Platen, Paul Jacob, Paul Wambergue, Paula Kurylowicz, Pavankumar Reddy Muddireddy, Philomène Chagniot, Pierre Stock, Pravesh Agrawal, Quentin Torroba, Romain Sauvestre, Roman Soletskyi, Rupert Menneer, Sagar Vaze, Samuel Barry, Sanchit Gandhi, Siddhant Waghjale, Siddharth Gandhi, Soham Ghosh, Srijan Mishra, Sumukh Aithal, Szymon Antoniak, Teven Le Scao, Théo Cachet, Theo Simon Sorg, Thibaut Lavril, Thiziri Nait Saada, Thomas Chabal, Thomas Foubert, Thomas Robert, Thomas Wang, Tim Lawson, Tom Bewley, Tom Bewley, Tom Edwards, Umar Jamil, Umberto Tomasini, Valeriia Nemychnikova, Van Phung, Vincent Maladière, Virgile Richard, Wassim Bouaziz, Wen-Ding Li, William Marshall, Xinghui Li, Xinyu Yang, Yassine El Ouahidi, Yihan Wang, Yunhao Tang, Zaccharie Ramzi</p><p><b>Upvotes:</b> 44</p><p><b>Summary:</b> We introduce the Ministral 3 series, a family of parameter-efficient dense language models designed for compute and memory constrained applications, available in three model sizes: 3B, 8B, and 14B parameters. For each model size, we release three variants: a pretrained base model for general-purpose use, an instruction finetuned, and a reasoning model for complex problem-solving. In addition, we present our recipe to derive the Ministral 3 models through Cascade Distillation, an iterative pruning and continued training with distillation technique. Each model comes with image understanding capabilities, all under the Apache 2.0 license.</p> - https://arxiv.org/abs/2601.08584 - Tue, 13 Jan 2026 14:06:03 +0000 + Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models + https://arxiv.org/abs/2601.14004 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14004.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hengyuan Zhang, Zhihao Zhang, Mingyang Wang, Zunhai Su, Yiwei Wang, Qianli Wang, Shuzhou Yuan, Ercong Nie, Xufeng Duan, Qibo Xue, Zeping Yu, Chenming Shang, Xiao Liang, Jing Xiong, Hui Shen, Chaofan Tao, Zhengwu Liu, Senjie Jin, Zhiheng Xi, Dongdong Zhang, Sophia Ananiadou, Tao Gui, Ruobing Xie, Hayden Kwok-Hay So, Hinrich Schütze, Xuanjing Huang, Qi Zhang, Ngai Wong</p><p><b>Upvotes:</b> 44</p><p><b>Summary:</b> Mechanistic Interpretability (MI) has emerged as a vital approach to demystify the opaque decision-making of Large Language Models (LLMs). However, existing reviews primarily treat MI as an observational science, summarizing analytical insights while lacking a systematic framework for actionable intervention. To bridge this gap, we present a practical survey structured around the pipeline: "Locate, Steer, and Improve." We formally categorize Localizing (diagnosis) and Steering (intervention) methods based on specific Interpretable Objects to establish a rigorous intervention protocol. Furthermore, we demonstrate how this framework enables tangible improvements in Alignment, Capability, and Efficiency, effectively operationalizing MI as an actionable methodology for model optimization. The curated paper list of this work is available at https://github.com/rattlesnakey/Awesome-Actionable-MI-Survey.</p> + https://arxiv.org/abs/2601.14004 + Tue, 20 Jan 2026 14:23:23 +0000 - End-to-End Video Character Replacement without Structural Guidance - https://arxiv.org/abs/2601.08587 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08587.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhengbo Xu, Jie Ma, Ziheng Wang, Zhan Peng, Jun Liang, Jing Li</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Controllable video character replacement with a user-provided identity remains a challenging problem due to the lack of paired video data. Prior works have predominantly relied on a reconstruction-based paradigm that requires per-frame segmentation masks and explicit structural guidance (e.g., skeleton, depth). This reliance, however, severely limits their generalizability in complex scenarios involving occlusions, character-object interactions, unusual poses, or challenging illumination, often leading to visual artifacts and temporal inconsistencies. In this paper, we propose MoCha, a pioneering framework that bypasses these limitations by requiring only a single arbitrary frame mask. To effectively adapt the multi-modal input condition and enhance facial identity, we introduce a condition-aware RoPE and employ an RL-based post-training stage. Furthermore, to overcome the scarcity of qualified paired-training data, we propose a comprehensive data construction pipeline. Specifically, we design three specialized datasets: a high-fidelity rendered dataset built with Unreal Engine 5 (UE5), an expression-driven dataset synthesized by current portrait animation techniques, and an augmented dataset derived from existing video-mask pairs. Extensive experiments demonstrate that our method substantially outperforms existing state-of-the-art approaches. We will release the code to facilitate further research. Please refer to our project page for more details: orange-3dv-team.github.io/MoCha</p> - https://arxiv.org/abs/2601.08587 - Tue, 13 Jan 2026 14:10:34 +0000 + Numina-Lean-Agent: An Open and General Agentic Reasoning System for Formal Mathematics + https://arxiv.org/abs/2601.14027 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14027.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Junqi Liu, Zihao Zhou, Zekai Zhu, Marco Dos Santos, Weikun He, Jiawei Liu, Ran Wang, Yunzhou Xie, Junqiao Zhao, Qiufeng Wang, Lihong Zhi, Jia Li, Wenda Li</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Agentic systems have recently become the dominant paradigm for formal theorem proving, achieving strong performance by coordinating multiple models and tools. However, existing approaches often rely on task-specific pipelines and trained formal provers, limiting their flexibility and reproducibility. In this paper, we propose the paradigm that directly uses a general coding agent as a formal math reasoner. This paradigm is motivated by (1) A general coding agent provides a natural interface for diverse reasoning tasks beyond proving, (2) Performance can be improved by simply replacing the underlying base model, without training, and (3) MCP enables flexible extension and autonomous calling of specialized tools, avoiding complex design. Based on this paradigm, we introduce Numina-Lean-Agent, which combines Claude Code with Numina-Lean-MCP to enable autonomous interaction with Lean, retrieval of relevant theorems, informal proving and auxiliary reasoning tools. Using Claude Opus 4.5 as the base model, Numina-Lean-Agent solves all problems in Putnam 2025 (12 / 12), matching the best closed-source system. Beyond benchmark evaluation, we further demonstrate its generality by interacting with mathematicians to successfully formalize the Brascamp-Lieb theorem. We release Numina-Lean-Agent and all solutions at https://github.com/project-numina/numina-lean-agent.</p> + https://arxiv.org/abs/2601.14027 + Tue, 20 Jan 2026 14:51:45 +0000 - ExpSeek: Self-Triggered Experience Seeking for Web Agents - https://arxiv.org/abs/2601.08605 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08605.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Wenyuan Zhang, Xinghua Zhang, Haiyang Yu, Shuaiyi Nie, Bingli Wu, Juwei Yue, Tingwen Liu, Yongbin Li</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Experience intervention in web agents emerges as a promising technical paradigm, enhancing agent interaction capabilities by providing valuable insights from accumulated experiences. However, existing methods predominantly inject experience passively as global context before task execution, struggling to adapt to dynamically changing contextual observations during agent-environment interaction. We propose ExpSeek, which shifts experience toward step-level proactive seeking: (1) estimating step-level entropy thresholds to determine intervention timing using the model's intrinsic signals; (2) designing step-level tailor-designed experience content. Experiments on Qwen3-8B and 32B models across four challenging web agent benchmarks demonstrate that ExpSeek achieves absolute improvements of 9.3% and 7.5%, respectively. Our experiments validate the feasibility and advantages of entropy as a self-triggering signal, reveal that even a 4B small-scale experience model can significantly boost the performance of larger agent models.</p> - https://arxiv.org/abs/2601.08605 - Tue, 13 Jan 2026 14:48:34 +0000 + PRiSM: Benchmarking Phone Realization in Speech Models + https://arxiv.org/abs/2601.14046 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14046.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shikhar Bharadwaj, Chin-Jou Li, Yoonjae Kim, Kwanghee Choi, Eunjung Yeo, Ryan Soh-Eun Shim, Hanyu Zhou, Brendon Boldt, Karen Rosero Jacome, Kalvin Chang, Darsh Agrawal, Keer Xu, Chao-Han Huck Yang, Jian Zhu, Shinji Watanabe, David R. Mortensen</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Phone recognition (PR) serves as the atomic interface for language-agnostic modeling for cross-lingual speech processing and phonetic analysis. Despite prolonged efforts in developing PR systems, current evaluations only measure surface-level transcription accuracy. We introduce PRiSM, the first open-source benchmark designed to expose blind spots in phonetic perception through intrinsic and extrinsic evaluation of PR systems. PRiSM standardizes transcription-based evaluation and assesses downstream utility in clinical, educational, and multilingual settings with transcription and representation probes. We find that diverse language exposure during training is key to PR performance, encoder-CTC models are the most stable, and specialized PR models still outperform Large Audio Language Models. PRiSM releases code, recipes, and datasets to move the field toward multilingual speech models with robust phonetic ability: https://github.com/changelinglab/prism.</p> + https://arxiv.org/abs/2601.14046 + Tue, 20 Jan 2026 15:00:36 +0000 - ViDoRe V3: A Comprehensive Evaluation of Retrieval Augmented Generation in Complex Real-World Scenarios - https://arxiv.org/abs/2601.08620 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08620.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> António Loison, Quentin Macé, Antoine Edy, Victor Xing, Tom Balough, Gabriel Moreira, Bo Liu, Manuel Faysse, Céline Hudelot, Gautier Viaud</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Retrieval-Augmented Generation (RAG) pipelines must address challenges beyond simple single-document retrieval, such as interpreting visual elements (tables, charts, images), synthesizing information across documents, and providing accurate source grounding. Existing benchmarks fail to capture this complexity, often focusing on textual data, single-document comprehension, or evaluating retrieval and generation in isolation. We introduce ViDoRe v3, a comprehensive multimodal RAG benchmark featuring multi-type queries over visually rich document corpora. It covers 10 datasets across diverse professional domains, comprising ~26,000 document pages paired with 3,099 human-verified queries, each available in 6 languages. Through 12,000 hours of human annotation effort, we provide high-quality annotations for retrieval relevance, bounding box localization, and verified reference answers. Our evaluation of state-of-the-art RAG pipelines reveals that visual retrievers outperform textual ones, late-interaction models and textual reranking substantially improve performance, and hybrid or purely visual contexts enhance answer generation quality. However, current models still struggle with non-textual elements, open-ended queries, and fine-grained visual grounding. To encourage progress in addressing these challenges, the benchmark is released under a commercially permissive license at https://hf.co/vidore.</p> - https://arxiv.org/abs/2601.08620 - Tue, 13 Jan 2026 15:00:33 +0000 + Lost in the Prompt Order: Revealing the Limitations of Causal Attention in Language Models + https://arxiv.org/abs/2601.14152 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14152.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hyunjong Ok, Jaeho Lee</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> Large language models exhibit surprising sensitivity to the structure of the prompt, but the mechanisms underlying this sensitivity remain poorly understood. In this work, we conduct an in-depth investigation on a striking case: in multiple-choice question answering, placing context before the questions and options (CQO) outperforms the reverse order (QOC) by over 14%p, consistently over a wide range of models and datasets. Through systematic architectural analysis, we identify causal attention as the core mechanism: in QOC prompts, the causal mask prevents option tokens from attending to context, creating an information bottleneck where context becomes invisible to options.</p> + https://arxiv.org/abs/2601.14152 + Tue, 20 Jan 2026 16:54:22 +0000 - VLingNav: Embodied Navigation with Adaptive Reasoning and Visual-Assisted Linguistic Memory - https://arxiv.org/abs/2601.08665 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08665.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shaoan Wang, Yuanfei Luo, Xingyu Chen, Aocheng Luo, Dongyue Li, Chang Liu, Sheng Chen, Yangang Zhang, Junzhi Yu</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> VLA models have shown promising potential in embodied navigation by unifying perception and planning while inheriting the strong generalization abilities of large VLMs. However, most existing VLA models rely on reactive mappings directly from observations to actions, lacking the explicit reasoning capabilities and persistent memory required for complex, long-horizon navigation tasks. To address these challenges, we propose VLingNav, a VLA model for embodied navigation grounded in linguistic-driven cognition. First, inspired by the dual-process theory of human cognition, we introduce an adaptive chain-of-thought mechanism, which dynamically triggers explicit reasoning only when necessary, enabling the agent to fluidly switch between fast, intuitive execution and slow, deliberate planning. Second, to handle long-horizon spatial dependencies, we develop a visual-assisted linguistic memory module that constructs a persistent, cross-modal semantic memory, enabling the agent to recall past observations to prevent repetitive exploration and infer movement trends for dynamic environments. For the training recipe, we construct Nav-AdaCoT-2.9M, the largest embodied navigation dataset with reasoning annotations to date, enriched with adaptive CoT annotations that induce a reasoning paradigm capable of adjusting both when to think and what to think about. Moreover, we incorporate an online expert-guided reinforcement learning stage, enabling the model to surpass pure imitation learning and to acquire more robust, self-explored navigation behaviors. Extensive experiments demonstrate that VLingNav achieves state-of-the-art performance across a wide range of embodied navigation benchmarks. Notably, VLingNav transfers to real-world robotic platforms in a zero-shot manner, executing various navigation tasks and demonstrating strong cross-domain and cross-task generalization.</p> - https://arxiv.org/abs/2601.08665 - Tue, 13 Jan 2026 15:43:43 +0000 + RoboBrain 2.5: Depth in Sight, Time in Mind + https://arxiv.org/abs/2601.14352 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14352.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Huajie Tan, Enshen Zhou, Zhiyu Li, Yijie Xu, Yuheng Ji, Xiansheng Chen, Cheng Chi, Pengwei Wang, Huizhu Jia, Yulong Ao, Mingyu Cao, Sixiang Chen, Zhe Li, Mengzhen Liu, Zixiao Wang, Shanyu Rong, Yaoxu Lyu, Zhongxia Zhao, Peterson Co, Yibo Li, Yi Han, Shaoxuan Xie, Guocai Yao, Songjing Wang, Leiduo Zhang, Xi Yang, Yance Jiao, Donghai Shi, Kunchang Xie, Shaokai Nie, Chunlei Men, Yonghua Lin, Zhongyuan Wang, Tiejun Huang, Shanghang Zhang</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> We introduce RoboBrain 2.5, a next-generation embodied AI foundation model that advances general perception, spatial reasoning, and temporal modeling through extensive training on high-quality spatiotemporal supervision. Building upon its predecessor, RoboBrain 2.5 introduces two major capability upgrades. Specifically, it unlocks Precise 3D Spatial Reasoning by shifting from 2D pixel-relative grounding to depth-aware coordinate prediction and absolute metric constraint comprehension, generating complete 3D manipulation traces as ordered keypoint sequences under physical constraints. Complementing this spatial precision, the model establishes Dense Temporal Value Estimation that provides dense, step-aware progress prediction and execution state understanding across varying viewpoints, producing stable feedback signals for downstream learning. Together, these upgrades extend the framework toward more physically grounded and execution-aware embodied intelligence for complex, fine-grained manipulation. The code and checkpoints are available at project website: https://superrobobrain.github.io</p> + https://arxiv.org/abs/2601.14352 + Tue, 20 Jan 2026 17:21:54 +0000 - Parallel Context-of-Experts Decoding for Retrieval Augmented Generation - https://arxiv.org/abs/2601.08670 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08670.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Giulio Corallo, Paolo Papotti</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> Retrieval Augmented Generation faces a trade-off: concatenating documents in a long prompt enables multi-document reasoning but creates prefill bottlenecks, while encoding document KV caches separately offers speed but breaks cross-document interaction. We propose Parallel Context-of-Experts Decoding (Pced), a training-free framework that shifts evidence aggregation from the attention mechanism to the decoding. Pced treats retrieved documents as isolated "experts", synchronizing their predictions via a novel retrieval-aware contrastive decoding rule that weighs expert logits against the model prior. This approach recovers cross-document reasoning capabilities without constructing a shared attention across documents.</p> - https://arxiv.org/abs/2601.08670 - Tue, 13 Jan 2026 15:46:59 +0000 + Paper2Rebuttal: A Multi-Agent Framework for Transparent Author Response Assistance + https://arxiv.org/abs/2601.14171 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14171.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Qianli Ma, Chang Guo, Zhiheng Tian, Siyu Wang, Jipeng Xiao, Yuanhao Yue, Zhipeng Zhang</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> Writing effective rebuttals is a high-stakes task that demands more than linguistic fluency, as it requires precise alignment between reviewer intent and manuscript details. Current solutions typically treat this as a direct-to-text generation problem, suffering from hallucination, overlooked critiques, and a lack of verifiable grounding. To address these limitations, we introduce RebuttalAgent, the first multi-agents framework that reframes rebuttal generation as an evidence-centric planning task. Our system decomposes complex feedback into atomic concerns and dynamically constructs hybrid contexts by synthesizing compressed summaries with high-fidelity text while integrating an autonomous and on-demand external search module to resolve concerns requiring outside literature. By generating an inspectable response plan before drafting, RebuttalAgent ensures that every argument is explicitly anchored in internal or external evidence. We validate our approach on the proposed RebuttalBench and demonstrate that our pipeline outperforms strong baselines in coverage, faithfulness, and strategic coherence, offering a transparent and controllable assistant for the peer review process. Code will be released.</p> + https://arxiv.org/abs/2601.14171 + Tue, 20 Jan 2026 17:23:51 +0000 - Rewarding the Rare: Uniqueness-Aware RL for Creative Problem Solving in LLMs - https://arxiv.org/abs/2601.08763 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08763.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhiyuan Hu, Yucheng Wang, Yufei He, Jiaying Wu, Yilun Zhao, See-Kiong Ng, Cynthia Breazeal, Anh Tuan Luu, Hae Won Park, Bryan Hooi</p><p><b>Upvotes:</b> 125</p><p><b>Summary:</b> Reinforcement learning (RL) has become a central paradigm for post-training large language models (LLMs), particularly for complex reasoning tasks, yet it often suffers from exploration collapse: policies prematurely concentrate on a small set of dominant reasoning patterns, improving pass@1 while limiting rollout-level diversity and gains in pass@k. We argue that this failure stems from regularizing local token behavior rather than diversity over sets of solutions. To address this, we propose Uniqueness-Aware Reinforcement Learning, a rollout-level objective that explicitly rewards correct solutions that exhibit rare high-level strategies. Our method uses an LLM-based judge to cluster rollouts for the same problem according to their high-level solution strategies, ignoring superficial variations, and reweights policy advantages inversely with cluster size. As a result, correct but novel strategies receive higher rewards than redundant ones. Across mathematics, physics, and medical reasoning benchmarks, our approach consistently improves pass@k across large sampling budgets and increases the area under the pass@k curve (AUC@K) without sacrificing pass@1, while sustaining exploration and uncovering more diverse solution strategies at scale.</p> - https://arxiv.org/abs/2601.08763 - Tue, 13 Jan 2026 17:48:43 +0000 + Toward Efficient Agents: Memory, Tool learning, and Planning + https://arxiv.org/abs/2601.14192 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14192.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xiaofang Yang, Lijun Li, Heng Zhou, Tong Zhu, Xiaoye Qu, Yuchen Fan, Qianshan Wei, Rui Ye, Li Kang, Yiran Qin, Zhiqiang Kou, Daizong Liu, Qi Li, Ning Ding, Siheng Chen, Jing Shao</p><p><b>Upvotes:</b> 49</p><p><b>Summary:</b> Recent years have witnessed increasing interest in extending large language models into agentic systems. While the effectiveness of agents has continued to improve, efficiency, which is crucial for real-world deployment, has often been overlooked. This paper therefore investigates efficiency from three core components of agents: memory, tool learning, and planning, considering costs such as latency, tokens, steps, etc. Aimed at conducting comprehensive research addressing the efficiency of the agentic system itself, we review a broad range of recent approaches that differ in implementation yet frequently converge on shared high-level principles including but not limited to bounding context via compression and management, designing reinforcement learning rewards to minimize tool invocation, and employing controlled search mechanisms to enhance efficiency, which we discuss in detail. Accordingly, we characterize efficiency in two complementary ways: comparing effectiveness under a fixed cost budget, and comparing cost at a comparable level of effectiveness. This trade-off can also be viewed through the Pareto frontier between effectiveness and cost. From this perspective, we also examine efficiency oriented benchmarks by summarizing evaluation protocols for these components and consolidating commonly reported efficiency metrics from both benchmark and methodological studies. Moreover, we discuss the key challenges and future directions, with the goal of providing promising insights.</p> + https://arxiv.org/abs/2601.14192 + Tue, 20 Jan 2026 17:51:56 +0000 - Motion Attribution for Video Generation - https://arxiv.org/abs/2601.08828 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08828.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xindi Wu, Despoina Paschalidou, Jun Gao, Antonio Torralba, Laura Leal-Taixé, Olga Russakovsky, Sanja Fidler, Jonathan Lorraine</p><p><b>Upvotes:</b> 65</p><p><b>Summary:</b> Despite the rapid progress of video generation models, the role of data in influencing motion is poorly understood. We present Motive (MOTIon attribution for Video gEneration), a motion-centric, gradient-based data attribution framework that scales to modern, large, high-quality video datasets and models. We use this to study which fine-tuning clips improve or degrade temporal dynamics. Motive isolates temporal dynamics from static appearance via motion-weighted loss masks, yielding efficient and scalable motion-specific influence computation. On text-to-video models, Motive identifies clips that strongly affect motion and guides data curation that improves temporal consistency and physical plausibility. With Motive-selected high-influence data, our method improves both motion smoothness and dynamic degree on VBench, achieving a 74.1% human preference win rate compared with the pretrained base model. To our knowledge, this is the first framework to attribute motion rather than visual appearance in video generative models and to use it to curate fine-tuning data.</p> - https://arxiv.org/abs/2601.08828 - Tue, 13 Jan 2026 18:59:09 +0000 + InT: Self-Proposed Interventions Enable Credit Assignment in LLM Reasoning + https://arxiv.org/abs/2601.14209 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14209.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Matthew Y. R. Yang, Hao Bai, Ian Wu, Gene Yang, Amrith Setlur, Aviral Kumar</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Outcome-reward reinforcement learning (RL) has proven effective at improving the reasoning capabilities of large language models (LLMs). However, standard RL assigns credit only at the level of the final answer, penalizing entire reasoning traces when the outcome is incorrect and uniformly reinforcing all steps when it is correct. As a result, correct intermediate steps may be discouraged in failed traces, while spurious steps may be reinforced in successful ones. We refer to this failure mode as the problem of credit assignment. While a natural remedy is to train a process reward model, accurately optimizing such models to identify corrective reasoning steps remains challenging. We introduce Intervention Training (InT), a training paradigm in which the model performs fine-grained credit assignment on its own reasoning traces by proposing short, targeted corrections that steer trajectories toward higher reward. Using reference solutions commonly available in mathematical reasoning datasets and exploiting the fact that verifying a model-generated solution is easier than generating a correct one from scratch, the model identifies the first error in its reasoning and proposes a single-step intervention to redirect the trajectory toward the correct solution. We then apply supervised fine-tuning (SFT) to the on-policy rollout up to the point of error concatenated with the intervention, localizing error to the specific step that caused failure. We show that the resulting model serves as a far better initialization for RL training. After running InT and subsequent fine-tuning with RL, we improve accuracy by nearly 14% over a 4B-parameter base model on IMO-AnswerBench, outperforming larger open-source models such as gpt-oss-20b.</p> + https://arxiv.org/abs/2601.14209 + Tue, 20 Jan 2026 18:15:38 +0000 - 3AM: Segment Anything with Geometric Consistency in Videos - https://arxiv.org/abs/2601.08831 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08831.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yang-Che Sun, Cheng Sun, Chin-Yang Lin, Fu-En Yang, Min-Hung Chen, Yen-Yu Lin, Yu-Lun Liu</p><p><b>Upvotes:</b> 33</p><p><b>Summary:</b> Video object segmentation methods like SAM2 achieve strong performance through memory-based architectures but struggle under large viewpoint changes due to reliance on appearance features. Traditional 3D instance segmentation methods address viewpoint consistency but require camera poses, depth maps, and expensive preprocessing. We introduce 3AM, a training-time enhancement that integrates 3D-aware features from MUSt3R into SAM2. Our lightweight Feature Merger fuses multi-level MUSt3R features that encode implicit geometric correspondence. Combined with SAM2's appearance features, the model achieves geometry-consistent recognition grounded in both spatial position and visual similarity. We propose a field-of-view aware sampling strategy ensuring frames observe spatially consistent object regions for reliable 3D correspondence learning. Critically, our method requires only RGB input at inference, with no camera poses or preprocessing. On challenging datasets with wide-baseline motion (ScanNet++, Replica), 3AM substantially outperforms SAM2 and extensions, achieving 90.6% IoU and 71.7% Positive IoU on ScanNet++'s Selected Subset, improving over state-of-the-art VOS methods by +15.9 and +30.4 points. Project page: https://jayisaking.github.io/3AM-Page/</p> - https://arxiv.org/abs/2601.08831 - Tue, 13 Jan 2026 18:59:54 +0000 + KAGE-Bench: Fast Known-Axis Visual Generalization Evaluation for Reinforcement Learning + https://arxiv.org/abs/2601.14232 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14232.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Egor Cherepanov, Daniil Zelezetsky, Alexey K. Kovalev, Aleksandr I. Panov</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Pixel-based reinforcement learning agents often fail under purely visual distribution shift even when latent dynamics and rewards are unchanged, but existing benchmarks entangle multiple sources of shift and hinder systematic analysis. We introduce KAGE-Env, a JAX-native 2D platformer that factorizes the observation process into independently controllable visual axes while keeping the underlying control problem fixed. By construction, varying a visual axis affects performance only through the induced state-conditional action distribution of a pixel policy, providing a clean abstraction for visual generalization. Building on this environment, we define KAGE-Bench, a benchmark of six known-axis suites comprising 34 train-evaluation configuration pairs that isolate individual visual shifts. Using a standard PPO-CNN baseline, we observe strong axis-dependent failures, with background and photometric shifts often collapsing success, while agent-appearance shifts are comparatively benign. Several shifts preserve forward motion while breaking task completion, showing that return alone can obscure generalization failures. Finally, the fully vectorized JAX implementation enables up to 33M environment steps per second on a single GPU, enabling fast and reproducible sweeps over visual factors. Code: https://avanturist322.github.io/KAGEBench/.</p> + https://arxiv.org/abs/2601.14232 + Tue, 20 Jan 2026 18:44:28 +0000 - Imagine-then-Plan: Agent Learning from Adaptive Lookahead with World Models - https://arxiv.org/abs/2601.08955 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.08955.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Youwei Liu, Jian Wang, Hanlin Wang, Beichen Guo, Wenjie Li</p><p><b>Upvotes:</b> 10</p><p><b>Summary:</b> Recent advances in world models have shown promise for modeling future dynamics of environmental states, enabling agents to reason and act without accessing real environments. Current methods mainly perform single-step or fixed-horizon rollouts, leaving their potential for complex task planning under-exploited. We propose Imagine-then-Plan (ITP), a unified framework for agent learning via lookahead imagination, where an agent's policy model interacts with the learned world model, yielding multi-step ``imagined'' trajectories. Since the imagination horizon may vary by tasks and stages, we introduce a novel adaptive lookahead mechanism by trading off the ultimate goal and task progress. The resulting imagined trajectories provide rich signals about future consequences, such as achieved progress and potential conflicts, which are fused with current observations, formulating a partially observable and imaginable Markov decision process to guide policy learning. We instantiate ITP with both training-free and reinforcement-trained variants. Extensive experiments across representative agent benchmarks demonstrate that ITP significantly outperforms competitive baselines. Further analyses validate that our adaptive lookahead largely enhances agents' reasoning capability, providing valuable insights into addressing broader, complex tasks.</p> - https://arxiv.org/abs/2601.08955 - Tue, 13 Jan 2026 19:49:58 +0000 + XR: Cross-Modal Agents for Composed Image Retrieval + https://arxiv.org/abs/2601.14245 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14245.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhongyu Yang, Wei Pang, Yingfang Yuan</p><p><b>Upvotes:</b> 9</p><p><b>Summary:</b> Retrieval is being redefined by agentic AI, demanding multimodal reasoning beyond conventional similarity-based paradigms. Composed Image Retrieval (CIR) exemplifies this shift as each query combines a reference image with textual modifications, requiring compositional understanding across modalities. While embedding-based CIR methods have achieved progress, they remain narrow in perspective, capturing limited cross-modal cues and lacking semantic reasoning. To address these limitations, we introduce XR, a training-free multi-agent framework that reframes retrieval as a progressively coordinated reasoning process. It orchestrates three specialized types of agents: imagination agents synthesize target representations through cross-modal generation, similarity agents perform coarse filtering via hybrid matching, and question agents verify factual consistency through targeted reasoning for fine filtering. Through progressive multi-agent coordination, XR iteratively refines retrieval to meet both semantic and visual query constraints, achieving up to a 38% gain over strong training-free and training-based baselines on FashionIQ, CIRR, and CIRCO, while ablations show each agent is essential. Code is available: https://01yzzyu.github.io/xr.github.io/.</p> + https://arxiv.org/abs/2601.14245 + Tue, 20 Jan 2026 18:57:00 +0000 - TranslateGemma Technical Report - https://arxiv.org/abs/2601.09012 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09012.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mara Finkelstein, Isaac Caswell, Tobias Domhan, Jan-Thorsten Peter, Juraj Juraska, Parker Riley, Daniel Deutsch, Cole Dilanni, Colin Cherry, Eleftheria Briakou, Elizabeth Nielsen, Jiaming Luo, Kat Black, Ryan Mullins, Sweta Agrawal, Wenda Xu, Erin Kats, Stephane Jaskiewicz, Markus Freitag, David Vilar</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> We present TranslateGemma, a suite of open machine translation models based on the Gemma 3 foundation models. To enhance the inherent multilingual capabilities of Gemma 3 for the translation task, we employ a two-stage fine-tuning process. First, supervised fine-tuning is performed using a rich mixture of high-quality large-scale synthetic parallel data generated via state-of-the-art models and human-translated parallel data. This is followed by a reinforcement learning phase, where we optimize translation quality using an ensemble of reward models, including MetricX-QE and AutoMQM, targeting translation quality. We demonstrate the effectiveness of TranslateGemma with human evaluation on the WMT25 test set across 10 language pairs and with automatic evaluation on the WMT24++ benchmark across 55 language pairs. Automatic metrics show consistent and substantial gains over the baseline Gemma 3 models across all sizes. Notably, smaller TranslateGemma models often achieve performance comparable to larger baseline models, offering improved efficiency. We also show that TranslateGemma models retain strong multimodal capabilities, with enhanced performance on the Vistra image translation benchmark. The release of the open TranslateGemma models aims to provide the research community with powerful and adaptable tools for machine translation.</p> - https://arxiv.org/abs/2601.09012 - Tue, 13 Jan 2026 22:23:24 +0000 + Which Reasoning Trajectories Teach Students to Reason Better? A Simple Metric of Informative Alignment + https://arxiv.org/abs/2601.14249 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14249.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yuming Yang, Mingyoung Lai, Wanxu Zhao, Xiaoran Fan, Zhiheng Xi, Mingqi Wu, Chiyue Huang, Jun Zhao, Haijun Lv, Jian Tong, Yunhua Zhou, Yicheng Zou, Qipeng Guo, Tao Gui, Qi Zhang, Xuanjing Huang</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Long chain-of-thought (CoT) trajectories provide rich supervision signals for distilling reasoning from teacher to student LLMs. However, both prior work and our experiments show that trajectories from stronger teachers do not necessarily yield better students, highlighting the importance of data-student suitability in distillation. Existing methods assess suitability primarily through student likelihood, favoring trajectories that closely align with the model's current behavior but overlooking more informative ones. Addressing this, we propose Rank-Surprisal Ratio (RSR), a simple metric that captures both alignment and informativeness to assess the suitability of a reasoning trajectory. RSR is motivated by the observation that effective trajectories typically combine low absolute probability with relatively high-ranked tokens under the student model, balancing learning signal strength and behavioral alignment. Concretely, RSR is defined as the ratio of a trajectory's average token-wise rank to its average negative log-likelihood, and is straightforward to compute and interpret. Across five student models and reasoning trajectories from 11 diverse teachers, RSR strongly correlates with post-training performance (average Spearman 0.86), outperforming existing metrics. We further demonstrate its practical utility in both trajectory selection and teacher selection.</p> + https://arxiv.org/abs/2601.14249 + Tue, 20 Jan 2026 18:58:10 +0000 - OpenDecoder: Open Large Language Model Decoding to Incorporate Document Quality in RAG - https://arxiv.org/abs/2601.09028 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09028.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Fengran Mo, Zhan Su, Yuchen Hui, Jinghan Zhang, Jia Ao Sun, Zheyuan Liu, Chao Zhang, Tetsuya Sakai, Jian-Yun Nie</p><p><b>Upvotes:</b> 30</p><p><b>Summary:</b> The development of large language models (LLMs) has achieved superior performance in a range of downstream tasks, including LLM-based retrieval-augmented generation (RAG). The quality of generated content heavily relies on the usefulness of the retrieved information and the capacity of LLMs' internal information processing mechanism to incorporate it in answer generation. It is generally assumed that the retrieved information is relevant to the question. However, the retrieved information may have a variable degree of relevance and usefulness, depending on the question and the document collection. It is important to take into account the relevance of the retrieved information in answer generation. In this paper, we propose OpenDecoder, a new approach that leverages explicit evaluation of the retrieved information as quality indicator features for generation. We aim to build a RAG model that is more robust to varying levels of noisy context. Three types of explicit evaluation information are considered: relevance score, ranking score, and QPP (query performance prediction) score. The experimental results on five benchmark datasets demonstrate the effectiveness and better robustness of OpenDecoder by outperforming various baseline methods. Importantly, this paradigm is flexible to be integrated with the post-training of LLMs for any purposes and incorporated with any type of external indicators.</p> - https://arxiv.org/abs/2601.09028 - Tue, 13 Jan 2026 23:26:30 +0000 + OmniTransfer: All-in-one Framework for Spatio-temporal Video Transfer + https://arxiv.org/abs/2601.14250 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14250.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Pengze Zhang, Yanze Wu, Mengtian Li, Xu Bai, Songtao Zhao, Fulong Ye, Chong Mou, Xinghui Li, Zhuowei Chen, Qian He, Mingyuan Gao</p><p><b>Upvotes:</b> 41</p><p><b>Summary:</b> Videos convey richer information than images or text, capturing both spatial and temporal dynamics. However, most existing video customization methods rely on reference images or task-specific temporal priors, failing to fully exploit the rich spatio-temporal information inherent in videos, thereby limiting flexibility and generalization in video generation. To address these limitations, we propose OmniTransfer, a unified framework for spatio-temporal video transfer. It leverages multi-view information across frames to enhance appearance consistency and exploits temporal cues to enable fine-grained temporal control. To unify various video transfer tasks, OmniTransfer incorporates three key designs: Task-aware Positional Bias that adaptively leverages reference video information to improve temporal alignment or appearance consistency; Reference-decoupled Causal Learning separating reference and target branches to enable precise reference transfer while improving efficiency; and Task-adaptive Multimodal Alignment using multimodal semantic guidance to dynamically distinguish and tackle different tasks. Extensive experiments show that OmniTransfer outperforms existing methods in appearance (ID and style) and temporal transfer (camera movement and video effects), while matching pose-guided methods in motion transfer without using pose, establishing a new paradigm for flexible, high-fidelity video generation.</p> + https://arxiv.org/abs/2601.14250 + Tue, 20 Jan 2026 18:58:11 +0000 - Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoning - https://arxiv.org/abs/2601.09088 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09088.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shaotian Yan, Kaiyuan Liu, Chen Shen, Bing Wang, Sinan Fan, Jun Zhang, Yue Wu, Zheng Wang, Jieping Ye</p><p><b>Upvotes:</b> 53</p><p><b>Summary:</b> In this report, we introduce DASD-4B-Thinking, a lightweight yet highly capable, fully open-source reasoning model. It achieves SOTA performance among open-source models of comparable scale across challenging benchmarks in mathematics, scientific reasoning, and code generation -- even outperforming several larger models. We begin by critically reexamining a widely adopted distillation paradigm in the community: SFT on teacher-generated responses, also known as sequence-level distillation. Although a series of recent works following this scheme have demonstrated remarkable efficiency and strong empirical performance, they are primarily grounded in the SFT perspective. Consequently, these approaches focus predominantly on designing heuristic rules for SFT data filtering, while largely overlooking the core principle of distillation itself -- enabling the student model to learn the teacher's full output distribution so as to inherit its generalization capability. Specifically, we identify three critical limitations in current practice: i) Inadequate representation of the teacher's sequence-level distribution; ii) Misalignment between the teacher's output distribution and the student's learning capacity; and iii) Exposure bias arising from teacher-forced training versus autoregressive inference. In summary, these shortcomings reflect a systemic absence of explicit teacher-student interaction throughout the distillation process, leaving the essence of distillation underexploited. To address these issues, we propose several methodological innovations that collectively form an enhanced sequence-level distillation training pipeline. Remarkably, DASD-4B-Thinking obtains competitive results using only 448K training samples -- an order of magnitude fewer than those employed by most existing open-source efforts. To support community research, we publicly release our models and the training dataset.</p> - https://arxiv.org/abs/2601.09088 - Wed, 14 Jan 2026 02:43:17 +0000 + LightOnOCR: A 1B End-to-End Multilingual Vision-Language Model for State-of-the-Art OCR + https://arxiv.org/abs/2601.14251 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14251.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Said Taghadouini, Adrien Cavaillès, Baptiste Aubertin</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> We present LightOnOCR-2-1B, a 1B-parameter end-to-end multilingual vision--language model that converts document images (e.g., PDFs) into clean, naturally ordered text without brittle OCR pipelines. Trained on a large-scale, high-quality distillation mix with strong coverage of scans, French documents, and scientific PDFs, LightOnOCR-2 achieves state-of-the-art results on OlmOCR-Bench while being 9times smaller and substantially faster than prior best-performing models. We further extend the output format to predict normalized bounding boxes for embedded images, introducing localization during pretraining via a resume strategy and refining it with RLVR using IoU-based rewards. Finally, we improve robustness with checkpoint averaging and task-arithmetic merging. We release model checkpoints under Apache 2.0, and publicly release the dataset and LightOnOCR-bbox-bench evaluation under their respective licenses.</p> + https://arxiv.org/abs/2601.14251 + Tue, 20 Jan 2026 18:58:32 +0000 - SkinFlow: Efficient Information Transmission for Open Dermatological Diagnosis via Dynamic Visual Encoding and Staged RL - https://arxiv.org/abs/2601.09136 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09136.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lijun Liu, Linwei Chen, Zhishou Zhang, Meng Tian, Hengfu Cui, Ruiyang Li, Zhaocheng Liu, Qiang Ju, Qianxi Li, Hong-Yu Zhou</p><p><b>Upvotes:</b> 35</p><p><b>Summary:</b> General-purpose Large Vision-Language Models (LVLMs), despite their massive scale, often falter in dermatology due to "diffuse attention" - the inability to disentangle subtle pathological lesions from background noise. In this paper, we challenge the assumption that parameter scaling is the only path to medical precision. We introduce SkinFlow, a framework that treats diagnosis as an optimization of visual information transmission efficiency. Our approach utilizes a Virtual-Width Dynamic Vision Encoder (DVE) to "unfold" complex pathological manifolds without physical parameter expansion, coupled with a two-stage Reinforcement Learning strategy. This strategy sequentially aligns explicit medical descriptions (Stage I) and reconstructs implicit diagnostic textures (Stage II) within a constrained semantic space. Furthermore, we propose a clinically grounded evaluation protocol that prioritizes diagnostic safety and hierarchical relevance over rigid label matching. Empirical results are compelling: our 7B model establishes a new state-of-the-art on the Fitzpatrick17k benchmark, achieving a +12.06% gain in Top-1 accuracy and a +28.57% boost in Top-6 accuracy over the massive general-purpose models (e.g., Qwen3VL-235B and GPT-5.2). These findings demonstrate that optimizing geometric capacity and information flow yields superior diagnostic reasoning compared to raw parameter scaling.</p> - https://arxiv.org/abs/2601.09136 - Wed, 14 Jan 2026 04:21:07 +0000 + Motion 3-to-4: 3D Motion Reconstruction for 4D Synthesis + https://arxiv.org/abs/2601.14253 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14253.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hongyuan Chen, Xingyu Chen, Youjia Zhang, Zexiang Xu, Anpei Chen</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> We present Motion 3-to-4, a feed-forward framework for synthesising high-quality 4D dynamic objects from a single monocular video and an optional 3D reference mesh. While recent advances have significantly improved 2D, video, and 3D content generation, 4D synthesis remains difficult due to limited training data and the inherent ambiguity of recovering geometry and motion from a monocular viewpoint. Motion 3-to-4 addresses these challenges by decomposing 4D synthesis into static 3D shape generation and motion reconstruction. Using a canonical reference mesh, our model learns a compact motion latent representation and predicts per-frame vertex trajectories to recover complete, temporally coherent geometry. A scalable frame-wise transformer further enables robustness to varying sequence lengths. Evaluations on both standard benchmarks and a new dataset with accurate ground-truth geometry show that Motion 3-to-4 delivers superior fidelity and spatial consistency compared to prior work. Project page is available at https://motion3-to-4.github.io/.</p> + https://arxiv.org/abs/2601.14253 + Tue, 20 Jan 2026 18:59:48 +0000 - EvasionBench: Detecting Evasive Answers in Financial Q&A via Multi-Model Consensus and LLM-as-Judge - https://arxiv.org/abs/2601.09142 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09142.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shijian Ma, Yan Lin, Yi Yang</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Detecting evasive answers in earnings calls is critical for financial transparency, yet progress is hindered by the lack of large-scale benchmarks. We introduce EvasionBench, comprising 30,000 training samples and 1,000 human-annotated test samples (Cohen's Kappa 0.835) across three evasion levels. Our key contribution is a multi-model annotation framework leveraging a core insight: disagreement between frontier LLMs signals hard examples most valuable for training. We mine boundary cases where two strong annotators conflict, using a judge to resolve labels. This approach outperforms single-model distillation by 2.4 percent, with judge-resolved samples improving generalization despite higher training loss (0.421 vs 0.393) - evidence that disagreement mining acts as implicit regularization. Our trained model Eva-4B (4B parameters) achieves 81.3 percent accuracy, outperforming its base by 25 percentage points and approaching frontier LLM performance at a fraction of inference cost.</p> - https://arxiv.org/abs/2601.09142 - Wed, 14 Jan 2026 04:26:43 +0000 + VideoMaMa: Mask-Guided Video Matting via Generative Prior + https://arxiv.org/abs/2601.14255 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14255.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Sangbeom Lim, Seoung Wug Oh, Jiahui Huang, Heeji Yoon, Seungryong Kim, Joon-Young Lee</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Generalizing video matting models to real-world videos remains a significant challenge due to the scarcity of labeled data. To address this, we present Video Mask-to-Matte Model (VideoMaMa) that converts coarse segmentation masks into pixel accurate alpha mattes, by leveraging pretrained video diffusion models. VideoMaMa demonstrates strong zero-shot generalization to real-world footage, even though it is trained solely on synthetic data. Building on this capability, we develop a scalable pseudo-labeling pipeline for large-scale video matting and construct the Matting Anything in Video (MA-V) dataset, which offers high-quality matting annotations for more than 50K real-world videos spanning diverse scenes and motions. To validate the effectiveness of this dataset, we fine-tune the SAM2 model on MA-V to obtain SAM2-Matte, which outperforms the same model trained on existing matting datasets in terms of robustness on in-the-wild videos. These findings emphasize the importance of large-scale pseudo-labeled video matting and showcase how generative priors and accessible segmentation cues can drive scalable progress in video matting research.</p> + https://arxiv.org/abs/2601.14255 + Tue, 20 Jan 2026 18:59:56 +0000 - Geometric Stability: The Missing Axis of Representations - https://arxiv.org/abs/2601.09173 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09173.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Prashant C. Raju</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Analysis of learned representations has a blind spot: it focuses on similarity, measuring how closely embeddings align with external references, but similarity reveals only what is represented, not whether that structure is robust. We introduce geometric stability, a distinct dimension that quantifies how reliably representational geometry holds under perturbation, and present Shesha, a framework for measuring it. Across 2,463 configurations in seven domains, we show that stability and similarity are empirically uncorrelated (ρapprox 0.01) and mechanistically distinct: similarity metrics collapse after removing the top principal components, while stability retains sensitivity to fine-grained manifold structure. This distinction yields actionable insights: for safety monitoring, stability acts as a functional geometric canary, detecting structural drift nearly 2times more sensitively than CKA while filtering out the non-functional noise that triggers false alarms in rigid distance metrics; for controllability, supervised stability predicts linear steerability (ρ= 0.89-0.96); for model selection, stability dissociates from transferability, revealing a geometric tax that transfer optimization incurs. Beyond machine learning, stability predicts CRISPR perturbation coherence and neural-behavioral coupling. By quantifying how reliably systems maintain structure, geometric stability provides a necessary complement to similarity for auditing representations across biological and computational systems.</p> - https://arxiv.org/abs/2601.09173 - Wed, 14 Jan 2026 05:15:22 +0000 + Implicit Neural Representation Facilitates Unified Universal Vision Encoding + https://arxiv.org/abs/2601.14256 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14256.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Matthew Gwilliam, Xiao Wang, Xuefeng Hu, Zhenheng Yang</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Models for image representation learning are typically designed for either recognition or generation. Various forms of contrastive learning help models learn to convert images to embeddings that are useful for classification, detection, and segmentation. On the other hand, models can be trained to reconstruct images with pixel-wise, perceptual, and adversarial losses in order to learn a latent space that is useful for image generation. We seek to unify these two directions with a first-of-its-kind model that learns representations which are simultaneously useful for recognition and generation. We train our model as a hyper-network for implicit neural representation, which learns to map images to model weights for fast, accurate reconstruction. We further integrate our INR hyper-network with knowledge distillation to improve its generalization and performance. Beyond the novel training design, the model also learns an unprecedented compressed embedding space with outstanding performance for various visual tasks. The complete model competes with state-of-the-art results for image representation learning, while also enabling generative capabilities with its high-quality tiny embeddings. The code is available at https://github.com/tiktok/huvr.</p> + https://arxiv.org/abs/2601.14256 + Tue, 20 Jan 2026 18:59:57 +0000 - MAXS: Meta-Adaptive Exploration with LLM Agents - https://arxiv.org/abs/2601.09259 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09259.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jian Zhang, Zhiyuan Wang, Zhangqi Wang, Yu He, Haoran Luo, li yuan, Lingling Zhang, Rui Mao, Qika Lin, Jun Liu</p><p><b>Upvotes:</b> 88</p><p><b>Summary:</b> Large Language Model (LLM) Agents exhibit inherent reasoning abilities through the collaboration of multiple tools. However, during agent inference, existing methods often suffer from (i) locally myopic generation, due to the absence of lookahead, and (ii) trajectory instability, where minor early errors can escalate into divergent reasoning paths. These issues make it difficult to balance global effectiveness and computational efficiency. To address these two issues, we propose meta-adaptive exploration with LLM agents https://github.com/exoskeletonzj/MAXS, a meta-adaptive reasoning framework based on LLM Agents that flexibly integrates tool execution and reasoning planning. MAXS employs a lookahead strategy to extend reasoning paths a few steps ahead, estimating the advantage value of tool usage, and combines step consistency variance and inter-step trend slopes to jointly select stable, consistent, and high-value reasoning steps. Additionally, we introduce a trajectory convergence mechanism that controls computational cost by halting further rollouts once path consistency is achieved, enabling a balance between resource efficiency and global effectiveness in multi-tool reasoning. We conduct extensive empirical studies across three base models (MiMo-VL-7B, Qwen2.5-VL-7B, Qwen2.5-VL-32B) and five datasets, demonstrating that MAXS consistently outperforms existing methods in both performance and inference efficiency. Further analysis confirms the effectiveness of our lookahead strategy and tool usage.</p> - https://arxiv.org/abs/2601.09259 - Wed, 14 Jan 2026 07:48:00 +0000 + Quantifying Speaker Embedding Phonological Rule Interactions in Accented Speech Synthesis + https://arxiv.org/abs/2601.14417 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14417.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Thanathai Lertpetchpun, Yoonjeong Lee, Thanapat Trachu, Jihwan Lee, Tiantian Feng, Dani Byrd, Shrikanth Narayanan</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Many spoken languages, including English, exhibit wide variation in dialects and accents, making accent control an important capability for flexible text-to-speech (TTS) models. Current TTS systems typically generate accented speech by conditioning on speaker embeddings associated with specific accents. While effective, this approach offers limited interpretability and controllability, as embeddings also encode traits such as timbre and emotion. In this study, we analyze the interaction between speaker embeddings and linguistically motivated phonological rules in accented speech synthesis. Using American and British English as a case study, we implement rules for flapping, rhoticity, and vowel correspondences. We propose the phoneme shift rate (PSR), a novel metric quantifying how strongly embeddings preserve or override rule-based transformations. Experiments show that combining rules with embeddings yields more authentic accents, while embeddings can attenuate or overwrite rules, revealing entanglement between accent and speaker identity. Our findings highlight rules as a lever for accent control and a framework for evaluating disentanglement in speech generation.</p> + https://arxiv.org/abs/2601.14417 + Tue, 20 Jan 2026 19:25:33 +0000 - A^3-Bench: Benchmarking Memory-Driven Scientific Reasoning via Anchor and Attractor Activation - https://arxiv.org/abs/2601.09274 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09274.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jian Zhang, Yu He, Zhiyuan Wang, Zhangqi Wang, Kai He, Fangzhi Xu, Qika Lin, Jun Liu</p><p><b>Upvotes:</b> 78</p><p><b>Summary:</b> Scientific reasoning relies not only on logical inference but also on activating prior knowledge and experiential structures. Memory can efficiently reuse knowledge and enhance reasoning consistency and stability. However, existing benchmarks mainly evaluate final answers or step-by-step coherence, overlooking the memory-driven mechanisms that underlie human reasoning, which involves activating anchors and attractors, then integrating them into multi-step inference. To address this gap, we propose A^3-Bench~ https://a3-bench.github.io, a benchmark designed to evaluate scientific reasoning through dual-scale memory-driven activation, grounded in Anchor and Attractor Activation. First, we annotate 2,198 science reasoning problems across domains using the SAPM process(subject, anchor & attractor, problem, and memory developing). Second, we introduce a dual-scale memory evaluation framework utilizing anchors and attractors, along with the AAUI(Anchor--Attractor Utilization Index) metric to measure memory activation rates. Finally, through experiments with various base models and paradigms, we validate A^3-Bench and analyze how memory activation impacts reasoning performance, providing insights into memory-driven scientific reasoning.</p> - https://arxiv.org/abs/2601.09274 - Wed, 14 Jan 2026 08:17:41 +0000 + GutenOCR: A Grounded Vision-Language Front-End for Documents + https://arxiv.org/abs/2601.14490 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14490.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hunter Heidenreich, Ben Elliott, Olivia Dinica, Yosheb Getachew</p><p><b>Upvotes:</b> 29</p><p><b>Summary:</b> GutenOCR is a family of grounded OCR front-ends obtained by fine-tuning Qwen2.5-VL-3B and Qwen2.5-VL-7B. The resulting single-checkpoint vision-language models expose reading, detection, and grounding through a unified, prompt-based interface. Trained on business documents, scientific articles, and synthetic grounding data, the models support full-page and localized reading with line- and paragraph-level bounding boxes and conditional ``where is x?'' queries. We introduce a grounded OCR evaluation protocol and show that GutenOCR-7B more than doubles the composite grounded OCR score of its Qwen2.5-VL-7B backbone on 10.5K held-out business and scientific pages (0.40 to 0.82). On Fox and OmniDocBench v1.5, our approach substantially improves region- and line-level OCR as well as text-detection recall, but reveals trade-offs in page-level linearization, color-guided OCR, and formula-heavy layouts.</p> + https://arxiv.org/abs/2601.14490 + Tue, 20 Jan 2026 21:26:15 +0000 - EvoFSM: Controllable Self-Evolution for Deep Research with Finite State Machines - https://arxiv.org/abs/2601.09465 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09465.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shuo Zhang, Chaofa Yuan, Ryan Guo, Xiaomin Yu, Rui Xu, Zhangquan Chen, Zinuo Li, Zhi Yang, Shuhao Guan, Zhenheng Tang, Sen Hu, Liwen Zhang, Ronghao Chen, Huacan Wang</p><p><b>Upvotes:</b> 38</p><p><b>Summary:</b> While LLM-based agents have shown promise for deep research, most existing approaches rely on fixed workflows that struggle to adapt to real-world, open-ended queries. Recent work therefore explores self-evolution by allowing agents to rewrite their own code or prompts to improve problem-solving ability, but unconstrained optimization often triggers instability, hallucinations, and instruction drift. We propose EvoFSM, a structured self-evolving framework that achieves both adaptability and control by evolving an explicit Finite State Machine (FSM) instead of relying on free-form rewriting. EvoFSM decouples the optimization space into macroscopic Flow (state-transition logic) and microscopic Skill (state-specific behaviors), enabling targeted improvements under clear behavioral boundaries. Guided by a critic mechanism, EvoFSM refines the FSM through a small set of constrained operations, and further incorporates a self-evolving memory that distills successful trajectories as reusable priors and failure patterns as constraints for future queries. Extensive evaluations on five multi-hop QA benchmarks demonstrate the effectiveness of EvoFSM. In particular, EvoFSM reaches 58.0% accuracy on the DeepSearch benchmark. Additional results on interactive decision-making tasks further validate its generalization.</p> - https://arxiv.org/abs/2601.09465 - Wed, 14 Jan 2026 13:19:13 +0000 + FARE: Fast-Slow Agentic Robotic Exploration + https://arxiv.org/abs/2601.14681 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14681.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shuhao Liao, Xuxin Lv, Jeric Lew, Shizhe Zhang, Jingsong Liang, Peizhuo Li, Yuhong Cao, Wenjun Wu, Guillaume Sartoretti</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> This work advances autonomous robot exploration by integrating agent-level semantic reasoning with fast local control. We introduce FARE, a hierarchical autonomous exploration framework that integrates a large language model (LLM) for global reasoning with a reinforcement learning (RL) policy for local decision making. FARE follows a fast-slow thinking paradigm. The slow-thinking LLM module interprets a concise textual description of the unknown environment and synthesizes an agent-level exploration strategy, which is then grounded into a sequence of global waypoints through a topological graph. To further improve reasoning efficiency, this module employs a modularity-based pruning mechanism that reduces redundant graph structures. The fast-thinking RL module executes exploration by reacting to local observations while being guided by the LLM-generated global waypoints. The RL policy is additionally shaped by a reward term that encourages adherence to the global waypoints, enabling coherent and robust closed-loop behavior. This architecture decouples semantic reasoning from geometric decision, allowing each module to operate in its appropriate temporal and spatial scale. In challenging simulated environments, our results show that FARE achieves substantial improvements in exploration efficiency over state-of-the-art baselines. We further deploy FARE on hardware and validate it in complex, large scale 200mtimes130m building environment.</p> + https://arxiv.org/abs/2601.14681 + Wed, 21 Jan 2026 05:56:24 +0000 - V-DPM: 4D Video Reconstruction with Dynamic Point Maps - https://arxiv.org/abs/2601.09499 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09499.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Edgar Sucar, Eldar Insafutdinov, Zihang Lai, Andrea Vedaldi</p><p><b>Upvotes:</b> 6</p><p><b>Summary:</b> Powerful 3D representations such as DUSt3R invariant point maps, which encode 3D shape and camera parameters, have significantly advanced feed forward 3D reconstruction. While point maps assume static scenes, Dynamic Point Maps (DPMs) extend this concept to dynamic 3D content by additionally representing scene motion. However, existing DPMs are limited to image pairs and, like DUSt3R, require post processing via optimization when more than two views are involved. We argue that DPMs are more useful when applied to videos and introduce V-DPM to demonstrate this. First, we show how to formulate DPMs for video input in a way that maximizes representational power, facilitates neural prediction, and enables reuse of pretrained models. Second, we implement these ideas on top of VGGT, a recent and powerful 3D reconstructor. Although VGGT was trained on static scenes, we show that a modest amount of synthetic data is sufficient to adapt it into an effective V-DPM predictor. Our approach achieves state of the art performance in 3D and 4D reconstruction for dynamic scenes. In particular, unlike recent dynamic extensions of VGGT such as P3, DPMs recover not only dynamic depth but also the full 3D motion of every point in the scene.</p> - https://arxiv.org/abs/2601.09499 - Wed, 14 Jan 2026 14:03:42 +0000 + Typhoon OCR: Open Vision-Language Model For Thai Document Extraction + https://arxiv.org/abs/2601.14722 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14722.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Surapon Nonesung, Natapong Nitarach, Teetouch Jaknamon, Pittawat Taveekitworachai, Kunat Pipatanakul</p><p><b>Upvotes:</b> 14</p><p><b>Summary:</b> Document extraction is a core component of digital workflows, yet existing vision-language models (VLMs) predominantly favor high-resource languages. Thai presents additional challenges due to script complexity from non-latin letters, the absence of explicit word boundaries, and the prevalence of highly unstructured real-world documents, limiting the effectiveness of current open-source models. This paper presents Typhoon OCR, an open VLM for document extraction tailored for Thai and English. The model is fine-tuned from vision-language backbones using a Thai-focused training dataset. The dataset is developed using a multi-stage data construction pipeline that combines traditional OCR, VLM-based restructuring, and curated synthetic data. Typhoon OCR is a unified framework capable of text transcription, layout reconstruction, and document-level structural consistency. The latest iteration of our model, Typhoon OCR V1.5, is a compact and inference-efficient model designed to reduce reliance on metadata and simplify deployment. Comprehensive evaluations across diverse Thai document categories, including financial reports, government forms, books, infographics, and handwritten documents, show that Typhoon OCR achieves performance comparable to or exceeding larger frontier proprietary models, despite substantially lower computational cost. The results demonstrate that open vision-language OCR models can achieve accurate text extraction and layout reconstruction for Thai documents, reaching performance comparable to proprietary systems while remaining lightweight and deployable.</p> + https://arxiv.org/abs/2601.14722 + Wed, 21 Jan 2026 07:24:32 +0000 - OpenVoxel: Training-Free Grouping and Captioning Voxels for Open-Vocabulary 3D Scene Understanding - https://arxiv.org/abs/2601.09575 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09575.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Sheng-Yu Huang, Jaesung Choe, Yu-Chiang Frank Wang, Cheng Sun</p><p><b>Upvotes:</b> 24</p><p><b>Summary:</b> We propose OpenVoxel, a training-free algorithm for grouping and captioning sparse voxels for the open-vocabulary 3D scene understanding tasks. Given the sparse voxel rasterization (SVR) model obtained from multi-view images of a 3D scene, our OpenVoxel is able to produce meaningful groups that describe different objects in the scene. Also, by leveraging powerful Vision Language Models (VLMs) and Multi-modal Large Language Models (MLLMs), our OpenVoxel successfully build an informative scene map by captioning each group, enabling further 3D scene understanding tasks such as open-vocabulary segmentation (OVS) or referring expression segmentation (RES). Unlike previous methods, our method is training-free and does not introduce embeddings from a CLIP/BERT text encoder. Instead, we directly proceed with text-to-text search using MLLMs. Through extensive experiments, our method demonstrates superior performance compared to recent studies, particularly in complex referring expression segmentation (RES) tasks. The code will be open.</p> - https://arxiv.org/abs/2601.09575 - Wed, 14 Jan 2026 15:45:57 +0000 + HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding + https://arxiv.org/abs/2601.14724 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14724.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Haowei Zhang, Shudong Yang, Jinlan Fu, See-Kiong Ng, Xipeng Qiu</p><p><b>Upvotes:</b> 67</p><p><b>Summary:</b> Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated significant improvement in offline video understanding. However, extending these capabilities to streaming video inputs, remains challenging, as existing models struggle to simultaneously maintain stable understanding performance, real-time responses, and low GPU memory overhead. To address this challenge, we propose HERMES, a novel training-free architecture for real-time and accurate understanding of video streams. Based on a mechanistic attention investigation, we conceptualize KV cache as a hierarchical memory framework that encapsulates video information across multiple granularities. During inference, HERMES reuses a compact KV cache, enabling efficient streaming understanding under resource constraints. Notably, HERMES requires no auxiliary computations upon the arrival of user queries, thereby guaranteeing real-time responses for continuous video stream interactions, which achieves 10times faster TTFT compared to prior SOTA. Even when reducing video tokens by up to 68% compared with uniform sampling, HERMES achieves superior or comparable accuracy across all benchmarks, with up to 11.4% gains on streaming datasets.</p> + https://arxiv.org/abs/2601.14724 + Wed, 21 Jan 2026 07:26:15 +0000 - Collaborative Multi-Agent Test-Time Reinforcement Learning for Reasoning - https://arxiv.org/abs/2601.09667 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09667.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zhiyuan Hu, Yunhai Hu, Juncheng Liu, Shuyue Stella Li, Yucheng Wang, Zhen Xu, See-Kiong Ng, Anh Tuan Luu, Xinxing Xu, Bryan Hooi, Cynthia Breazeal, Hae Won Park</p><p><b>Upvotes:</b> 73</p><p><b>Summary:</b> Multi-agent systems have evolved into practical LLM-driven collaborators for many applications, gaining robustness from diversity and cross-checking. However, multi-agent RL (MARL) training is resource-intensive and unstable: co-adapting teammates induce non-stationarity, and rewards are often sparse and high-variance. Therefore, we introduce Multi-Agent Test-Time Reinforcement Learning (MATTRL), a framework that injects structured textual experience into multi-agent deliberation at inference time. MATTRL forms a multi-expert team of specialists for multi-turn discussions, retrieves and integrates test-time experiences, and reaches consensus for final decision-making. We also study credit assignment for constructing a turn-level experience pool, then reinjecting it into the dialogue. Across challenging benchmarks in medicine, math, and education, MATTRL improves accuracy by an average of 3.67\% over a multi-agent baseline, and by 8.67\% over comparable single-agent baselines. Ablation studies examine different credit-assignment schemes and provide a detailed comparison of how they affect training outcomes. MATTRL offers a stable, effective and efficient path to distribution-shift-robust multi-agent reasoning without tuning.</p> - https://arxiv.org/abs/2601.09667 - Wed, 14 Jan 2026 17:57:43 +0000 + Render-of-Thought: Rendering Textual Chain-of-Thought as Images for Visual Latent Reasoning + https://arxiv.org/abs/2601.14750 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.14750.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yifan Wang, Shiyu Li, Peiming Li, Xiaochen Yang, Yang Tang, Zheng Wei</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Chain-of-Thought (CoT) prompting has achieved remarkable success in unlocking the reasoning capabilities of Large Language Models (LLMs). Although CoT prompting enhances reasoning, its verbosity imposes substantial computational overhead. Recent works often focus exclusively on outcome alignment and lack supervision on the intermediate reasoning process. These deficiencies obscure the analyzability of the latent reasoning chain. To address these challenges, we introduce Render-of-Thought (RoT), the first framework to reify the reasoning chain by rendering textual steps into images, making the latent rationale explicit and traceable. Specifically, we leverage the vision encoders of existing Vision Language Models (VLMs) as semantic anchors to align the vision embeddings with the textual space. This design ensures plug-and-play implementation without incurring additional pre-training overhead. Extensive experiments on mathematical and logical reasoning benchmarks demonstrate that our method achieves 3-4x token compression and substantial inference acceleration compared to explicit CoT. Furthermore, it maintains competitive performance against other methods, validating the feasibility of this paradigm. Our code is available at https://github.com/TencentBAC/RoT</p> + https://arxiv.org/abs/2601.14750 + Wed, 21 Jan 2026 08:09:25 +0000 - STEP3-VL-10B Technical Report - https://arxiv.org/abs/2601.09668 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09668.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ailin Huang, Chengyuan Yao, Chunrui Han, Fanqi Wan, Hangyu Guo, Haoran Lv, Hongyu Zhou, Jia Wang, Jian Zhou, Jianjian Sun, Jingcheng Hu, Kangheng Lin, Liang Zhao, Mitt Huang, Song Yuan, Wenwen Qu, Xiangfeng Wang, Yanlin Lai, Yingxiu Zhao, Yinmin Zhang, Yukang Shi, Yuyang Chen, Zejia Weng, Ziyang Meng, Ang Li, Aobo Kong, Bo Dong, Changyi Wan, David Wang, Di Qi, Dingming Li, En Yu, Guopeng Li, Haiquan Yin, Han Zhou, Hanshan Zhang, Haolong Yan, Hebin Zhou, Hongbo Peng, Jiaran Zhang, Jiashu Lv, Jiayi Fu, Jie Cheng, Jie Zhou, Jisheng Yin, Jingjing Xie, Jingwei Wu, Jun Zhang, Junfeng Liu, Kaijun Tan, Kaiwen Yan, Liangyu Chen, Lina Chen, Mingliang Li, Qian Zhao, Quan Sun, Shaoliang Pang, Shengjie Fan, Shijie Shang, Siyuan Zhang, Tianhao You, Wei Ji, Wuxun Xie, Xiaobo Yang, Xiaojie Hou, Xiaoran Jiao, Xiaoxiao Ren, Xiangwen Kong, Xin Huang, Xin Wu, Xing Chen, Xinran Wang, Xuelin Zhang, Yana Wei, Yang Li, Yanming Xu, Yeqing Shen, Yuang Peng, Yue Peng, Yu Zhou, Yusheng Li, Yuxiang Yang, Yuyang Zhang, Zhe Xie, Zhewei Huang, Zhenyi Lu, Zhimin Fan, Zihui Cheng, Daxin Jiang, Qi Han, Xiangyu Zhang, Yibo Zhu, Zheng Ge</p><p><b>Upvotes:</b> 161</p><p><b>Summary:</b> We present STEP3-VL-10B, a lightweight open-source foundation model designed to redefine the trade-off between compact efficiency and frontier-level multimodal intelligence. STEP3-VL-10B is realized through two strategic shifts: first, a unified, fully unfrozen pre-training strategy on 1.2T multimodal tokens that integrates a language-aligned Perception Encoder with a Qwen3-8B decoder to establish intrinsic vision-language synergy; and second, a scaled post-training pipeline featuring over 1k iterations of reinforcement learning. Crucially, we implement Parallel Coordinated Reasoning (PaCoRe) to scale test-time compute, allocating resources to scalable perceptual reasoning that explores and synthesizes diverse visual hypotheses. Consequently, despite its compact 10B footprint, STEP3-VL-10B rivals or surpasses models 10times-20times larger (e.g., GLM-4.6V-106B, Qwen3-VL-235B) and top-tier proprietary flagships like Gemini 2.5 Pro and Seed-1.5-VL. Delivering best-in-class performance, it records 92.2% on MMBench and 80.11% on MMMU, while excelling in complex reasoning with 94.43% on AIME2025 and 75.95% on MathVision. We release the full model suite to provide the community with a powerful, efficient, and reproducible baseline.</p> - https://arxiv.org/abs/2601.09668 - Wed, 14 Jan 2026 17:58:24 +0000 + The Responsibility Vacuum: Organizational Failure in Scaled Agent Systems + https://arxiv.org/abs/2601.15059 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15059.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Oleg Romanchuk, Roman Bondar</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> Modern CI/CD pipelines integrating agent-generated code exhibit a structural failure in responsibility attribution. Decisions are executed through formally correct approval processes, yet no entity possesses both the authority to approve those decisions and the epistemic capacity to meaningfully understand their basis. We define this condition as responsibility vacuum: a state in which decisions occur, but responsibility cannot be attributed because authority and verification capacity do not coincide. We show that this is not a process deviation or technical defect, but a structural property of deployments where decision generation throughput exceeds bounded human verification capacity. We identify a scaling limit under standard deployment assumptions, including parallel agent generation, CI-based validation, and individualized human approval gates. Beyond a throughput threshold, verification ceases to function as a decision criterion and is replaced by ritualized approval based on proxy signals. Personalized responsibility becomes structurally unattainable in this regime. We further characterize a CI amplification dynamic, whereby increasing automated validation coverage raises proxy signal density without restoring human capacity. Under fixed time and attention constraints, this accelerates cognitive offloading in the broad sense and widens the gap between formal approval and epistemic understanding. Additional automation therefore amplifies, rather than mitigates, the responsibility vacuum. We conclude that unless organizations explicitly redesign decision boundaries or reassign responsibility away from individual decisions toward batch- or system-level ownership, responsibility vacuum remains an invisible but persistent failure mode in scaled agent deployments.</p> + https://arxiv.org/abs/2601.15059 + Wed, 21 Jan 2026 15:05:27 +0000 - DeepResearchEval: An Automated Framework for Deep Research Task Construction and Agentic Evaluation - https://arxiv.org/abs/2601.09688 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09688.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yibo Wang, Lei Wang, Yue Deng, Keming Wu, Yao Xiao, Huanjin Yao, Liwei Kang, Hai Ye, Yongcheng Jing, Lidong Bing</p><p><b>Upvotes:</b> 109</p><p><b>Summary:</b> Deep research systems are widely used for multi-step web research, analysis, and cross-source synthesis, yet their evaluation remains challenging. Existing benchmarks often require annotation-intensive task construction, rely on static evaluation dimensions, or fail to reliably verify facts when citations are missing. To bridge these gaps, we introduce DeepResearchEval, an automated framework for deep research task construction and agentic evaluation. For task construction, we propose a persona-driven pipeline generating realistic, complex research tasks anchored in diverse user profiles, applying a two-stage filter Task Qualification and Search Necessity to retain only tasks requiring multi-source evidence integration and external retrieval. For evaluation, we propose an agentic pipeline with two components: an Adaptive Point-wise Quality Evaluation that dynamically derives task-specific evaluation dimensions, criteria, and weights conditioned on each generated task, and an Active Fact-Checking that autonomously extracts and verifies report statements via web search, even when citations are missing.</p> - https://arxiv.org/abs/2601.09688 - Wed, 14 Jan 2026 18:38:31 +0000 + Facilitating Proactive and Reactive Guidance for Decision Making on the Web: A Design Probe with WebSeek + https://arxiv.org/abs/2601.15100 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15100.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yanwei Huang, Arpit Narechania</p><p><b>Upvotes:</b> 3</p><p><b>Summary:</b> Web AI agents such as ChatGPT Agent and GenSpark are increasingly used for routine web-based tasks, yet they still rely on text-based input prompts, lack proactive detection of user intent, and offer no support for interactive data analysis and decision making. We present WebSeek, a mixed-initiative browser extension that enables users to discover and extract information from webpages to then flexibly build, transform, and refine tangible data artifacts-such as tables, lists, and visualizations-all within an interactive canvas. Within this environment, users can perform analysis-including data transformations such as joining tables or creating visualizations-while an in-built AI both proactively offers context-aware guidance and automation, and reactively responds to explicit user requests. An exploratory user study (N=15) with WebSeek as a probe reveals participants' diverse analysis strategies, underscoring their desire for transparency and control during human-AI collaboration.</p> + https://arxiv.org/abs/2601.15100 + Wed, 21 Jan 2026 15:38:57 +0000 - Efficient Camera-Controlled Video Generation of Static Scenes via Sparse Diffusion and 3D Rendering - https://arxiv.org/abs/2601.09697 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09697.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jieying Chen, Jeffrey Hu, Joan Lasenby, Ayush Tewari</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Modern video generative models based on diffusion models can produce very realistic clips, but they are computationally inefficient, often requiring minutes of GPU time for just a few seconds of video. This inefficiency poses a critical barrier to deploying generative video in applications that require real-time interactions, such as embodied AI and VR/AR. This paper explores a new strategy for camera-conditioned video generation of static scenes: using diffusion-based generative models to generate a sparse set of keyframes, and then synthesizing the full video through 3D reconstruction and rendering. By lifting keyframes into a 3D representation and rendering intermediate views, our approach amortizes the generation cost across hundreds of frames while enforcing geometric consistency. We further introduce a model that predicts the optimal number of keyframes for a given camera trajectory, allowing the system to adaptively allocate computation. Our final method, SRENDER, uses very sparse keyframes for simple trajectories and denser ones for complex camera motion. This results in video generation that is more than 40 times faster than the diffusion-based baseline in generating 20 seconds of video, while maintaining high visual fidelity and temporal stability, offering a practical path toward efficient and controllable video synthesis.</p> - https://arxiv.org/abs/2601.09697 - Wed, 14 Jan 2026 18:50:06 +0000 + The Flexibility Trap: Why Arbitrary Order Limits Reasoning Potential in Diffusion Language Models + https://arxiv.org/abs/2601.15165 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15165.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Zanlin Ni, Shenzhi Wang, Yang Yue, Tianyu Yu, Weilin Zhao, Yeguo Hua, Tianyi Chen, Jun Song, Cheng Yu, Bo Zheng, Gao Huang</p><p><b>Upvotes:</b> 59</p><p><b>Summary:</b> Diffusion Large Language Models (dLLMs) break the rigid left-to-right constraint of traditional LLMs, enabling token generation in arbitrary orders. Intuitively, this flexibility implies a solution space that strictly supersets the fixed autoregressive trajectory, theoretically unlocking superior reasoning potential for general tasks like mathematics and coding. Consequently, numerous works have leveraged reinforcement learning (RL) to elicit the reasoning capability of dLLMs. In this paper, we reveal a counter-intuitive reality: arbitrary order generation, in its current form, narrows rather than expands the reasoning boundary of dLLMs. We find that dLLMs tend to exploit this order flexibility to bypass high-uncertainty tokens that are crucial for exploration, leading to a premature collapse of the solution space. This observation challenges the premise of existing RL approaches for dLLMs, where considerable complexities, such as handling combinatorial trajectories and intractable likelihoods, are often devoted to preserving this flexibility. We demonstrate that effective reasoning is better elicited by intentionally forgoing arbitrary order and applying standard Group Relative Policy Optimization (GRPO) instead. Our approach, JustGRPO, is minimalist yet surprisingly effective (e.g., 89.1% accuracy on GSM8K) while fully retaining the parallel decoding ability of dLLMs. Project page: https://nzl-thu.github.io/the-flexibility-trap</p> + https://arxiv.org/abs/2601.15165 + Wed, 21 Jan 2026 16:41:58 +0000 - Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning - https://arxiv.org/abs/2601.09708 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09708.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chi-Pin Huang, Yunze Man, Zhiding Yu, Min-Hung Chen, Jan Kautz, Yu-Chiang Frank Wang, Fu-En Yang</p><p><b>Upvotes:</b> 48</p><p><b>Summary:</b> Vision-Language-Action (VLA) tasks require reasoning over complex visual scenes and executing adaptive actions in dynamic environments. While recent studies on reasoning VLAs show that explicit chain-of-thought (CoT) can improve generalization, they suffer from high inference latency due to lengthy reasoning traces. We propose Fast-ThinkAct, an efficient reasoning framework that achieves compact yet performant planning through verbalizable latent reasoning. Fast-ThinkAct learns to reason efficiently with latent CoTs by distilling from a teacher, driven by a preference-guided objective to align manipulation trajectories that transfers both linguistic and visual planning capabilities for embodied control. This enables reasoning-enhanced policy learning that effectively connects compact reasoning to action execution. Extensive experiments across diverse embodied manipulation and reasoning benchmarks demonstrate that Fast-ThinkAct achieves strong performance with up to 89.3\% reduced inference latency over state-of-the-art reasoning VLAs, while maintaining effective long-horizon planning, few-shot adaptation, and failure recovery.</p> - https://arxiv.org/abs/2601.09708 - Wed, 14 Jan 2026 18:59:59 +0000 + BayesianVLA: Bayesian Decomposition of Vision Language Action Models via Latent Action Queries + https://arxiv.org/abs/2601.15197 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15197.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shijie Lian, Bin Yu, Xiaopeng Lin, Laurence T. Yang, Zhaolong Shen, Changti Wu, Yuzhuo Miao, Cong Huang, Kai Chen</p><p><b>Upvotes:</b> 53</p><p><b>Summary:</b> Vision-Language-Action (VLA) models have shown promise in robot manipulation but often struggle to generalize to new instructions or complex multi-task scenarios. We identify a critical pathology in current training paradigms where goal-driven data collection creates a dataset bias. In such datasets, language instructions are highly predictable from visual observations alone, causing the conditional mutual information between instructions and actions to vanish, a phenomenon we term Information Collapse. Consequently, models degenerate into vision-only policies that ignore language constraints and fail in out-of-distribution (OOD) settings. To address this, we propose BayesianVLA, a novel framework that enforces instruction following via Bayesian decomposition. By introducing learnable Latent Action Queries, we construct a dual-branch architecture to estimate both a vision-only prior p(a mid v) and a language-conditioned posterior π(a mid v, ell). We then optimize the policy to maximize the conditional Pointwise Mutual Information (PMI) between actions and instructions. This objective effectively penalizes the vision shortcut and rewards actions that explicitly explain the language command. Without requiring new data, BayesianVLA significantly improves generalization. Extensive experiments across on SimplerEnv and RoboCasa demonstrate substantial gains, including an 11.3% improvement on the challenging OOD SimplerEnv benchmark, validating the ability of our approach to robustly ground language in action.</p> + https://arxiv.org/abs/2601.15197 + Wed, 21 Jan 2026 17:15:22 +0000 - Transition Matching Distillation for Fast Video Generation - https://arxiv.org/abs/2601.09881 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.09881.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Weili Nie, Julius Berner, Nanye Ma, Chao Liu, Saining Xie, Arash Vahdat</p><p><b>Upvotes:</b> 22</p><p><b>Summary:</b> Large video diffusion and flow models have achieved remarkable success in high-quality video generation, but their use in real-time interactive applications remains limited due to their inefficient multi-step sampling process. In this work, we present Transition Matching Distillation (TMD), a novel framework for distilling video diffusion models into efficient few-step generators. The central idea of TMD is to match the multi-step denoising trajectory of a diffusion model with a few-step probability transition process, where each transition is modeled as a lightweight conditional flow. To enable efficient distillation, we decompose the original diffusion backbone into two components: (1) a main backbone, comprising the majority of early layers, that extracts semantic representations at each outer transition step; and (2) a flow head, consisting of the last few layers, that leverages these representations to perform multiple inner flow updates. Given a pretrained video diffusion model, we first introduce a flow head to the model, and adapt it into a conditional flow map. We then apply distribution matching distillation to the student model with flow head rollout in each transition step. Extensive experiments on distilling Wan2.1 1.3B and 14B text-to-video models demonstrate that TMD provides a flexible and strong trade-off between generation speed and visual quality. In particular, TMD outperforms existing distilled models under comparable inference costs in terms of visual fidelity and prompt adherence. Project page: https://research.nvidia.com/labs/genair/tmd</p> - https://arxiv.org/abs/2601.09881 - Wed, 14 Jan 2026 21:30:03 +0000 + Privacy Collapse: Benign Fine-Tuning Can Break Contextual Privacy in Language Models + https://arxiv.org/abs/2601.15220 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15220.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Anmol Goel, Cornelius Emde, Sangdoo Yun, Seong Joon Oh, Martin Gubri</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> We identify a novel phenomenon in language models: benign fine-tuning of frontier models can lead to privacy collapse. We find that diverse, subtle patterns in training data can degrade contextual privacy, including optimisation for helpfulness, exposure to user information, emotional and subjective dialogue, and debugging code printing internal variables, among others. Fine-tuned models lose their ability to reason about contextual privacy norms, share information inappropriately with tools, and violate memory boundaries across contexts. Privacy collapse is a ``silent failure'' because models maintain high performance on standard safety and utility benchmarks whilst exhibiting severe privacy vulnerabilities. Our experiments show evidence of privacy collapse across six models (closed and open weight), five fine-tuning datasets (real-world and controlled data), and two task categories (agentic and memory-based). Our mechanistic analysis reveals that privacy representations are uniquely fragile to fine-tuning, compared to task-relevant features which are preserved. Our results reveal a critical gap in current safety evaluations, in particular for the deployment of specialised agents.</p> + https://arxiv.org/abs/2601.15220 + Wed, 21 Jan 2026 17:53:06 +0000 - CoF-T2I: Video Models as Pure Visual Reasoners for Text-to-Image Generation - https://arxiv.org/abs/2601.10061 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10061.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chengzhuo Tong, Mingkun Chang, Shenglong Zhang, Yuran Wang, Cheng Liang, Zhizheng Zhao, Ruichuan An, Bohan Zeng, Yang Shi, Yifan Dai, Ziming Zhao, Guanbin Li, Pengfei Wan, Yuanxing Zhang, Wentao Zhang</p><p><b>Upvotes:</b> 28</p><p><b>Summary:</b> Recent video generation models have revealed the emergence of Chain-of-Frame (CoF) reasoning, enabling frame-by-frame visual inference. With this capability, video models have been successfully applied to various visual tasks (e.g., maze solving, visual puzzles). However, their potential to enhance text-to-image (T2I) generation remains largely unexplored due to the absence of a clearly defined visual reasoning starting point and interpretable intermediate states in the T2I generation process. To bridge this gap, we propose CoF-T2I, a model that integrates CoF reasoning into T2I generation via progressive visual refinement, where intermediate frames act as explicit reasoning steps and the final frame is taken as output. To establish such an explicit generation process, we curate CoF-Evol-Instruct, a dataset of CoF trajectories that model the generation process from semantics to aesthetics. To further improve quality and avoid motion artifacts, we enable independent encoding operation for each frame. Experiments show that CoF-T2I significantly outperforms the base video model and achieves competitive performance on challenging benchmarks, reaching 0.86 on GenEval and 7.468 on Imagine-Bench. These results indicate the substantial promise of video models for advancing high-quality text-to-image generation.</p> - https://arxiv.org/abs/2601.10061 - Thu, 15 Jan 2026 04:33:06 +0000 + PROGRESSLM: Towards Progress Reasoning in Vision-Language Models + https://arxiv.org/abs/2601.15224 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15224.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jianshu Zhang, Chengxuan Qian, Haosen Sun, Haoran Lu, Dingcheng Wang, Letian Xue, Han Liu</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Estimating task progress requires reasoning over long-horizon dynamics rather than recognizing static visual content. While modern Vision-Language Models (VLMs) excel at describing what is visible, it remains unclear whether they can infer how far a task has progressed from partial observations. To this end, we introduce Progress-Bench, a benchmark for systematically evaluating progress reasoning in VLMs. Beyond benchmarking, we further explore a human-inspired two-stage progress reasoning paradigm through both training-free prompting and training-based approach based on curated dataset ProgressLM-45K. Experiments on 14 VLMs show that most models are not yet ready for task progress estimation, exhibiting sensitivity to demonstration modality and viewpoint changes, as well as poor handling of unanswerable cases. While training-free prompting that enforces structured progress reasoning yields limited and model-dependent gains, the training-based ProgressLM-3B achieves consistent improvements even at a small model scale, despite being trained on a task set fully disjoint from the evaluation tasks. Further analyses reveal characteristic error patterns and clarify when and why progress reasoning succeeds or fails.</p> + https://arxiv.org/abs/2601.15224 + Wed, 21 Jan 2026 17:56:59 +0000 - Deriving Character Logic from Storyline as Codified Decision Trees - https://arxiv.org/abs/2601.10080 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10080.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Letian Peng, Kun Zhou, Longfei Yun, Yupeng Hou, Jingbo Shang</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> Role-playing (RP) agents rely on behavioral profiles to act consistently across diverse narrative contexts, yet existing profiles are largely unstructured, non-executable, and weakly validated, leading to brittle agent behavior. We propose Codified Decision Trees (CDT), a data-driven framework that induces an executable and interpretable decision structure from large-scale narrative data. CDT represents behavioral profiles as a tree of conditional rules, where internal nodes correspond to validated scene conditions and leaves encode grounded behavioral statements, enabling deterministic retrieval of context-appropriate rules at execution time. The tree is learned by iteratively inducing candidate scene-action rules, validating them against data, and refining them through hierarchical specialization, yielding profiles that support transparent inspection and principled updates. Across multiple benchmarks, CDT substantially outperforms human-written profiles and prior profile induction methods on 85 characters across 16 artifacts, indicating that codified and validated behavioral representations lead to more reliable agent grounding.</p> - https://arxiv.org/abs/2601.10080 - Thu, 15 Jan 2026 05:12:43 +0000 + OpenVision 3: A Family of Unified Visual Encoder for Both Understanding and Generation + https://arxiv.org/abs/2601.15369 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15369.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Letian Zhang, Sucheng Ren, Yanqing Liu, Xianhang Li, Zeyu Wang, Yuyin Zhou, Huaxiu Yao, Zeyu Zheng, Weili Nie, Guilin Liu, Zhiding Yu, Cihang Xie</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> This paper presents a family of advanced vision encoder, named OpenVision 3, that learns a single, unified visual representation that can serve both image understanding and image generation. Our core architecture is simple: we feed VAE-compressed image latents to a ViT encoder and train its output to support two complementary roles. First, the encoder output is passed to the ViT-VAE decoder to reconstruct the original image, encouraging the representation to capture generative structure. Second, the same representation is optimized with contrastive learning and image-captioning objectives, strengthening semantic features. By jointly optimizing reconstruction- and semantics-driven signals in a shared latent space, the encoder learns representations that synergize and generalize well across both regimes. We validate this unified design through extensive downstream evaluations with the encoder frozen. For multimodal understanding, we plug the encoder into the LLaVA-1.5 framework: it performs comparably with a standard CLIP vision encoder (e.g., 62.4 vs 62.2 on SeedBench, and 83.7 vs 82.9 on POPE). For generation, we test it under the RAE framework: ours substantially surpasses the standard CLIP-based encoder (e.g., gFID: 1.89 vs 2.54 on ImageNet). We hope this work can spur future research on unified modeling.</p> + https://arxiv.org/abs/2601.15369 + Wed, 21 Jan 2026 18:47:12 +0000 - FlowAct-R1: Towards Interactive Humanoid Video Generation - https://arxiv.org/abs/2601.10103 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10103.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Lizhen Wang, Yongming Zhu, Zhipeng Ge, Youwei Zheng, Longhao Zhang, Tianshu Hu, Shiyang Qin, Mingshuang Luo, Jiaxu Zhang, Xin Chen, Yulong Wang, Zerong Zheng, Jianwen Jiang, Chao Liang, Weifeng Chen, Xing Wang, Yuan Zhang, Mingyuan Gao</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Interactive humanoid video generation aims to synthesize lifelike visual agents that can engage with humans through continuous and responsive video. Despite recent advances in video synthesis, existing methods often grapple with the trade-off between high-fidelity synthesis and real-time interaction requirements. In this paper, we propose FlowAct-R1, a framework specifically designed for real-time interactive humanoid video generation. Built upon a MMDiT architecture, FlowAct-R1 enables the streaming synthesis of video with arbitrary durations while maintaining low-latency responsiveness. We introduce a chunkwise diffusion forcing strategy, complemented by a novel self-forcing variant, to alleviate error accumulation and ensure long-term temporal consistency during continuous interaction. By leveraging efficient distillation and system-level optimizations, our framework achieves a stable 25fps at 480p resolution with a time-to-first-frame (TTFF) of only around 1.5 seconds. The proposed method provides holistic and fine-grained full-body control, enabling the agent to transition naturally between diverse behavioral states in interactive scenarios. Experimental results demonstrate that FlowAct-R1 achieves exceptional behavioral vividness and perceptual realism, while maintaining robust generalization across diverse character styles.</p> - https://arxiv.org/abs/2601.10103 - Thu, 15 Jan 2026 06:16:22 +0000 + Rethinking Video Generation Model for the Embodied World + https://arxiv.org/abs/2601.15282 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15282.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yufan Deng, Zilin Pan, Hongyu Zhang, Xiaojie Li, Ruoqing Hu, Yufei Ding, Yiming Zou, Yan Zeng, Daquan Zhou</p><p><b>Upvotes:</b> 42</p><p><b>Summary:</b> Video generation models have significantly advanced embodied intelligence, unlocking new possibilities for generating diverse robot data that capture perception, reasoning, and action in the physical world. However, synthesizing high-quality videos that accurately reflect real-world robotic interactions remains challenging, and the lack of a standardized benchmark limits fair comparisons and progress. To address this gap, we introduce a comprehensive robotics benchmark, RBench, designed to evaluate robot-oriented video generation across five task domains and four distinct embodiments. It assesses both task-level correctness and visual fidelity through reproducible sub-metrics, including structural consistency, physical plausibility, and action completeness. Evaluation of 25 representative models highlights significant deficiencies in generating physically realistic robot behaviors. Furthermore, the benchmark achieves a Spearman correlation coefficient of 0.96 with human evaluations, validating its effectiveness. While RBench provides the necessary lens to identify these deficiencies, achieving physical realism requires moving beyond evaluation to address the critical shortage of high-quality training data. Driven by these insights, we introduce a refined four-stage data pipeline, resulting in RoVid-X, the largest open-source robotic dataset for video generation with 4 million annotated video clips, covering thousands of tasks and enriched with comprehensive physical property annotations. Collectively, this synergistic ecosystem of evaluation and data establishes a robust foundation for rigorous assessment and scalable training of video models, accelerating the evolution of embodied AI toward general intelligence.</p> + https://arxiv.org/abs/2601.15282 + Wed, 21 Jan 2026 18:59:18 +0000 - LaViT: Aligning Latent Visual Thoughts for Multi-modal Reasoning - https://arxiv.org/abs/2601.10129 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10129.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Linquan Wu, Tianxiang Jiang, Yifei Dong, Haoyu Yang, Fengji Zhang, Shichaang Meng, Ai Xuan, Linqi Song, Jacky Keung</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Current multimodal latent reasoning often relies on external supervision (e.g., auxiliary images), ignoring intrinsic visual attention dynamics. In this work, we identify a critical Perception Gap in distillation: student models frequently mimic a teacher's textual output while attending to fundamentally divergent visual regions, effectively relying on language priors rather than grounded perception. To bridge this, we propose LaViT, a framework that aligns latent visual thoughts rather than static embeddings. LaViT compels the student to autoregressively reconstruct the teacher's visual semantics and attention trajectories prior to text generation, employing a curriculum sensory gating mechanism to prevent shortcut learning. Extensive experiments show that LaViT significantly enhances visual grounding, achieving up to +16.9% gains on complex reasoning tasks and enabling a compact 3B model to outperform larger open-source variants and proprietary models like GPT-4o.</p> - https://arxiv.org/abs/2601.10129 - Thu, 15 Jan 2026 07:14:24 +0000 + VIOLA: Towards Video In-Context Learning with Minimal Annotations + https://arxiv.org/abs/2601.15549 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15549.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ryo Fujii, Hideo Saito, Ryo Hachiuma</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> Generalizing Multimodal Large Language Models (MLLMs) to novel video domains is essential for real-world deployment but remains challenging due to the scarcity of labeled data. While In-Context Learning (ICL) offers a training-free adaptation path, standard methods rely on large annotated pools, which are often impractical in specialized environments like industrial or surgical settings since they require the experts' annotations. To bridge this gap, we introduce VIOLA (Video In-cOntext Learning with minimal Annotation), a label-efficient framework that synergizes minimal expert supervision with abundant unlabeled data. First, to maximize the efficiency of a strict annotation budget, we propose density-uncertainty-weighted sampling. Unlike standard diversity or uncertainty strategies that risk selecting visual outliers, our method leverages density estimation to identify samples that are simultaneously diverse, representative, and informative. Second, to utilize the remaining unlabeled data without noise propagation, we construct a hybrid pool and introduce confidence-aware retrieval and confidence-aware prompting. These mechanisms explicitly model label reliability, retrieving demonstrations based on a composite score of similarity and confidence while enabling the MLLM to adaptively distinguish between verified ground truths and noisy pseudo-labels. Extensive experiments across nine diverse benchmarks using four MLLMs demonstrate that our framework significantly outperforms various baselines in low-resource settings, achieving robust adaptation with minimal annotation costs.</p> + https://arxiv.org/abs/2601.15549 + Thu, 22 Jan 2026 00:35:30 +0000 - M^4olGen: Multi-Agent, Multi-Stage Molecular Generation under Precise Multi-Property Constraints - https://arxiv.org/abs/2601.10131 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10131.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yizhan Li, Florence Cloutier, Sifan Wu, Ali Parviz, Boris Knyazev, Yan Zhang, Glen Berseth, Bang Liu</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Generating molecules that satisfy precise numeric constraints over multiple physicochemical properties is critical and challenging. Although large language models (LLMs) are expressive, they struggle with precise multi-objective control and numeric reasoning without external structure and feedback. We introduce M olGen, a fragment-level, retrieval-augmented, two-stage framework for molecule generation under multi-property constraints. Stage I : Prototype generation: a multi-agent reasoner performs retrieval-anchored, fragment-level edits to produce a candidate near the feasible region. Stage II : RL-based fine-grained optimization: a fragment-level optimizer trained with Group Relative Policy Optimization (GRPO) applies one- or multi-hop refinements to explicitly minimize the property errors toward our target while regulating edit complexity and deviation from the prototype. A large, automatically curated dataset with reasoning chains of fragment edits and measured property deltas underpins both stages, enabling deterministic, reproducible supervision and controllable multi-hop reasoning. Unlike prior work, our framework better reasons about molecules by leveraging fragments and supports controllable refinement toward numeric targets. Experiments on generation under two sets of property constraints (QED, LogP, Molecular Weight and HOMO, LUMO) show consistent gains in validity and precise satisfaction of multi-property targets, outperforming strong LLMs and graph-based algorithms.</p> - https://arxiv.org/abs/2601.10131 - Thu, 15 Jan 2026 07:18:05 +0000 + Qwen3-TTS Technical Report + https://arxiv.org/abs/2601.15621 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15621.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hangrui Hu, Xinfa Zhu, Ting He, Dake Guo, Bin Zhang, Xiong Wang, Zhifang Guo, Ziyue Jiang, Hongkun Hao, Zishan Guo, Xinyu Zhang, Pei Zhang, Baosong Yang, Jin Xu, Jingren Zhou, Junyang Lin</p><p><b>Upvotes:</b> 32</p><p><b>Summary:</b> In this report, we present the Qwen3-TTS series, a family of advanced multilingual, controllable, robust, and streaming text-to-speech models. Qwen3-TTS supports state-of-the-art 3-second voice cloning and description-based control, allowing both the creation of entirely novel voices and fine-grained manipulation over the output speech. Trained on over 5 million hours of speech data spanning 10 languages, Qwen3-TTS adopts a dual-track LM architecture for real-time synthesis, coupled with two speech tokenizers: 1) Qwen-TTS-Tokenizer-25Hz is a single-codebook codec emphasizing semantic content, which offers seamlessly integration with Qwen-Audio and enables streaming waveform reconstruction via a block-wise DiT. 2) Qwen-TTS-Tokenizer-12Hz achieves extreme bitrate reduction and ultra-low-latency streaming, enabling immediate first-packet emission (97,ms) through its 12.5 Hz, 16-layer multi-codebook design and a lightweight causal ConvNet. Extensive experiments indicate state-of-the-art performance across diverse objective and subjective benchmark (e.g., TTS multilingual test set, InstructTTSEval, and our long speech test set). To facilitate community research and development, we release both tokenizers and models under the Apache 2.0 license.</p> + https://arxiv.org/abs/2601.15621 + Thu, 22 Jan 2026 03:51:43 +0000 - ToolSafe: Enhancing Tool Invocation Safety of LLM-based agents via Proactive Step-level Guardrail and Feedback - https://arxiv.org/abs/2601.10156 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10156.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yutao Mou, Zhangchi Xue, Lijun Li, Peiyang Liu, Shikun Zhang, Wei Ye, Jing Shao</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> While LLM-based agents can interact with environments via invoking external tools, their expanded capabilities also amplify security risks. Monitoring step-level tool invocation behaviors in real time and proactively intervening before unsafe execution is critical for agent deployment, yet remains under-explored. In this work, we first construct TS-Bench, a novel benchmark for step-level tool invocation safety detection in LLM agents. We then develop a guardrail model, TS-Guard, using multi-task reinforcement learning. The model proactively detects unsafe tool invocation actions before execution by reasoning over the interaction history. It assesses request harmfulness and action-attack correlations, producing interpretable and generalizable safety judgments and feedback. Furthermore, we introduce TS-Flow, a guardrail-feedback-driven reasoning framework for LLM agents, which reduces harmful tool invocations of ReAct-style agents by 65 percent on average and improves benign task completion by approximately 10 percent under prompt injection attacks.</p> - https://arxiv.org/abs/2601.10156 - Thu, 15 Jan 2026 07:54:32 +0000 + From Passive Metric to Active Signal: The Evolving Role of Uncertainty Quantification in Large Language Models + https://arxiv.org/abs/2601.15690 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15690.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaxin Zhang, Wendi Cui, Zhuohang Li, Lifu Huang, Bradley Malin, Caiming Xiong, Chien-Sheng Wu</p><p><b>Upvotes:</b> 4</p><p><b>Summary:</b> While Large Language Models (LLMs) show remarkable capabilities, their unreliability remains a critical barrier to deployment in high-stakes domains. This survey charts a functional evolution in addressing this challenge: the evolution of uncertainty from a passive diagnostic metric to an active control signal guiding real-time model behavior. We demonstrate how uncertainty is leveraged as an active control signal across three frontiers: in advanced reasoning to optimize computation and trigger self-correction; in autonomous agents to govern metacognitive decisions about tool use and information seeking; and in reinforcement learning to mitigate reward hacking and enable self-improvement via intrinsic rewards. By grounding these advancements in emerging theoretical frameworks like Bayesian methods and Conformal Prediction, we provide a unified perspective on this transformative trend. This survey provides a comprehensive overview, critical analysis, and practical design patterns, arguing that mastering the new trend of uncertainty is essential for building the next generation of scalable, reliable, and trustworthy AI.</p> + https://arxiv.org/abs/2601.15690 + Thu, 22 Jan 2026 06:21:31 +0000 - PRL: Process Reward Learning Improves LLMs' Reasoning Ability and Broadens the Reasoning Boundary - https://arxiv.org/abs/2601.10201 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10201.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiarui Yao, Ruida Wang, Tong Zhang</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Improving the reasoning abilities of Large Language Models (LLMs) has been a continuous topic recently. But most relevant works are based on outcome rewards at the trajectory level, missing fine-grained supervision during the reasoning process. Other existing training frameworks that try to combine process signals together to optimize LLMs also rely heavily on tedious additional steps like MCTS, training a separate reward model, etc., doing harm to the training efficiency. Moreover, the intuition behind the process signals design lacks rigorous theoretical support, leaving the understanding of the optimization mechanism opaque. In this paper, we propose Process Reward Learning (PRL), which decomposes the entropy regularized reinforcement learning objective into intermediate steps, with rigorous process rewards that could be assigned to models accordingly. Starting from theoretical motivation, we derive the formulation of PRL that is essentially equivalent to the objective of reward maximization plus a KL-divergence penalty term between the policy model and a reference model. However, PRL could turn the outcome reward into process supervision signals, which helps better guide the exploration during RL optimization. From our experiment results, we demonstrate that PRL not only improves the average performance for LLMs' reasoning ability measured by average @ n, but also broadens the reasoning boundary by improving the pass @ n metric. Extensive experiments show the effectiveness of PRL could be verified and generalized.</p> - https://arxiv.org/abs/2601.10201 - Thu, 15 Jan 2026 09:01:53 +0000 + Agentic Uncertainty Quantification + https://arxiv.org/abs/2601.15703 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15703.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaxin Zhang, Prafulla Kumar Choubey, Kung-Hsiang Huang, Caiming Xiong, Chien-Sheng Wu</p><p><b>Upvotes:</b> 8</p><p><b>Summary:</b> Although AI agents have demonstrated impressive capabilities in long-horizon reasoning, their reliability is severely hampered by the ``Spiral of Hallucination,'' where early epistemic errors propagate irreversibly. Existing methods face a dilemma: uncertainty quantification (UQ) methods typically act as passive sensors, only diagnosing risks without addressing them, while self-reflection mechanisms suffer from continuous or aimless corrections. To bridge this gap, we propose a unified Dual-Process Agentic UQ (AUQ) framework that transforms verbalized uncertainty into active, bi-directional control signals. Our architecture comprises two complementary mechanisms: System 1 (Uncertainty-Aware Memory, UAM), which implicitly propagates verbalized confidence and semantic explanations to prevent blind decision-making; and System 2 (Uncertainty-Aware Reflection, UAR), which utilizes these explanations as rational cues to trigger targeted inference-time resolution only when necessary. This enables the agent to balance efficient execution and deep deliberation dynamically. Extensive experiments on closed-loop benchmarks and open-ended deep research tasks demonstrate that our training-free approach achieves superior performance and trajectory-level calibration. We believe this principled framework AUQ represents a significant step towards reliable agents.</p> + https://arxiv.org/abs/2601.15703 + Thu, 22 Jan 2026 07:16:26 +0000 - DanQing: An Up-to-Date Large-Scale Chinese Vision-Language Pre-training Dataset - https://arxiv.org/abs/2601.10305 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10305.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Hengyu Shen, Tiancheng Gu, Bin Qin, Lan Wu, Yuling Wu, Shuo Tan, Zelong Sun, Jun Wang, Nan Wu, Xiang An, Weidong Cai, Ziyong Feng, Kaicheng Yang</p><p><b>Upvotes:</b> 32</p><p><b>Summary:</b> Vision-Language Pre-training (VLP) models demonstrate strong performance across various downstream tasks by learning from large-scale image-text pairs through contrastive pretraining. The release of extensive English image-text datasets (e.g., COYO-700M and LAION-400M) has enabled widespread adoption of models such as CLIP and SigLIP in tasks including cross-modal retrieval and image captioning. However, the advancement of Chinese vision-language pretraining has substantially lagged behind, due to the scarcity of high-quality Chinese image-text data. To address this gap, we develop a comprehensive pipeline for constructing a high-quality Chinese cross-modal dataset. As a result, we propose DanQing, which contains 100 million image-text pairs collected from Common Crawl. Different from existing datasets, DanQing is curated through a more rigorous selection process, yielding superior data quality. Moreover, DanQing is primarily built from 2024-2025 web data, enabling models to better capture evolving semantic trends and thus offering greater practical utility. We compare DanQing with existing datasets by continual pre-training of the SigLIP2 model. Experimental results show that DanQing consistently achieves superior performance across a range of Chinese downstream tasks, including zero-shot classification, cross-modal retrieval, and LMM-based evaluations. To facilitate further research in Chinese vision-language pre-training, we will open-source the DanQing dataset under the Creative Common CC-BY 4.0 license.</p> - https://arxiv.org/abs/2601.10305 - Thu, 15 Jan 2026 11:28:58 +0000 + Towards Automated Kernel Generation in the Era of LLMs + https://arxiv.org/abs/2601.15727 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15727.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yang Yu, Peiyu Zang, Chi Hsu Tsai, Haiming Wu, Yixin Shen, Jialing Zhang, Haoyu Wang, Zhiyou Xiao, Jingze Shi, Yuyu Luo, Wentao Zhang, Chunlei Men, Guang Liu, Yonghua Lin</p><p><b>Upvotes:</b> 16</p><p><b>Summary:</b> The performance of modern AI systems is fundamentally constrained by the quality of their underlying kernels, which translate high-level algorithmic semantics into low-level hardware operations. Achieving near-optimal kernels requires expert-level understanding of hardware architectures and programming models, making kernel engineering a critical but notoriously time-consuming and non-scalable process. Recent advances in large language models (LLMs) and LLM-based agents have opened new possibilities for automating kernel generation and optimization. LLMs are well-suited to compress expert-level kernel knowledge that is difficult to formalize, while agentic systems further enable scalable optimization by casting kernel development as an iterative, feedback-driven loop. Rapid progress has been made in this area. However, the field remains fragmented, lacking a systematic perspective for LLM-driven kernel generation. This survey addresses this gap by providing a structured overview of existing approaches, spanning LLM-based approaches and agentic optimization workflows, and systematically compiling the datasets and benchmarks that underpin learning and evaluation in this domain. Moreover, key open challenges and future research directions are further outlined, aiming to establish a comprehensive reference for the next generation of automated kernel optimization. To keep track of this field, we maintain an open-source GitHub repository at https://github.com/flagos-ai/awesome-LLM-driven-kernel-generation.</p> + https://arxiv.org/abs/2601.15727 + Thu, 22 Jan 2026 07:53:52 +0000 - Think-Then-Generate: Reasoning-Aware Text-to-Image Diffusion with LLM Encoders - https://arxiv.org/abs/2601.10332 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10332.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Siqi Kou, Jiachun Jin, Zetong Zhou, Ye Ma, Yugang Wang, Quan Chen, Peng Jiang, Xiao Yang, Jun Zhu, Kai Yu, Zhijie Deng</p><p><b>Upvotes:</b> 25</p><p><b>Summary:</b> Recent progress in text-to-image (T2I) diffusion models (DMs) has enabled high-quality visual synthesis from diverse textual prompts. Yet, most existing T2I DMs, even those equipped with large language model (LLM)-based text encoders, remain text-pixel mappers -- they employ LLMs merely as text encoders, without leveraging their inherent reasoning capabilities to infer what should be visually depicted given the textual prompt. To move beyond such literal generation, we propose the think-then-generate (T2G) paradigm, where the LLM-based text encoder is encouraged to reason about and rewrite raw user prompts; the states of the rewritten prompts then serve as diffusion conditioning. To achieve this, we first activate the think-then-rewrite pattern of the LLM encoder with a lightweight supervised fine-tuning process. Subsequently, the LLM encoder and diffusion backbone are co-optimized to ensure faithful reasoning about the context and accurate rendering of the semantics via Dual-GRPO. In particular, the text encoder is reinforced using image-grounded rewards to infer and recall world knowledge, while the diffusion backbone is pushed to produce semantically consistent and visually coherent images. Experiments show substantial improvements in factual consistency, semantic alignment, and visual realism across reasoning-based image generation and editing benchmarks, achieving 0.79 on WISE score, nearly on par with GPT-4. Our results constitute a promising step toward next-generation unified models with reasoning, expression, and demonstration capacities.</p> - https://arxiv.org/abs/2601.10332 - Thu, 15 Jan 2026 12:19:05 +0000 + Agentic Confidence Calibration + https://arxiv.org/abs/2601.15778 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15778.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jiaxin Zhang, Caiming Xiong, Chien-Sheng Wu</p><p><b>Upvotes:</b> 5</p><p><b>Summary:</b> AI agents are rapidly advancing from passive language models to autonomous systems executing complex, multi-step tasks. Yet their overconfidence in failure remains a fundamental barrier to deployment in high-stakes settings. Existing calibration methods, built for static single-turn outputs, cannot address the unique challenges of agentic systems, such as compounding errors along trajectories, uncertainty from external tools, and opaque failure modes. To address these challenges, we introduce, for the first time, the problem of Agentic Confidence Calibration and propose Holistic Trajectory Calibration (HTC), a novel diagnostic framework that extracts rich process-level features ranging from macro dynamics to micro stability across an agent's entire trajectory. Powered by a simple, interpretable model, HTC consistently surpasses strong baselines in both calibration and discrimination, across eight benchmarks, multiple LLMs, and diverse agent frameworks. Beyond performance, HTC delivers three essential advances: it provides interpretability by revealing the signals behind failure, enables transferability by applying across domains without retraining, and achieves generalization through a General Agent Calibrator (GAC) that achieves the best calibration (lowest ECE) on the out-of-domain GAIA benchmark. Together, these contributions establish a new process-centric paradigm for confidence calibration, providing a framework for diagnosing and enhancing the reliability of AI agents.</p> + https://arxiv.org/abs/2601.15778 + Thu, 22 Jan 2026 09:08:25 +0000 - Toward Ultra-Long-Horizon Agentic Science: Cognitive Accumulation for Machine Learning Engineering - https://arxiv.org/abs/2601.10402 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10402.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xinyu Zhu, Yuzhu Cai, Zexi Liu, Bingyang Zheng, Cheng Wang, Rui Ye, Jiaao Chen, Hanrui Wang, Wei-Chen Wang, Yuzhi Zhang, Linfeng Zhang, Weinan E, Di Jin, Siheng Chen</p><p><b>Upvotes:</b> 34</p><p><b>Summary:</b> The advancement of artificial intelligence toward agentic science is currently bottlenecked by the challenge of ultra-long-horizon autonomy, the ability to sustain strategic coherence and iterative correction over experimental cycles spanning days or weeks. While Large Language Models (LLMs) have demonstrated prowess in short-horizon reasoning, they are easily overwhelmed by execution details in the high-dimensional, delayed-feedback environments of real-world research, failing to consolidate sparse feedback into coherent long-term guidance. Here, we present ML-Master 2.0, an autonomous agent that masters ultra-long-horizon machine learning engineering (MLE) which is a representative microcosm of scientific discovery. By reframing context management as a process of cognitive accumulation, our approach introduces Hierarchical Cognitive Caching (HCC), a multi-tiered architecture inspired by computer systems that enables the structural differentiation of experience over time. By dynamically distilling transient execution traces into stable knowledge and cross-task wisdom, HCC allows agents to decouple immediate execution from long-term experimental strategy, effectively overcoming the scaling limits of static context windows. In evaluations on OpenAI's MLE-Bench under 24-hour budgets, ML-Master 2.0 achieves a state-of-the-art medal rate of 56.44%. Our findings demonstrate that ultra-long-horizon autonomy provides a scalable blueprint for AI capable of autonomous exploration beyond human-precedent complexities.</p> - https://arxiv.org/abs/2601.10402 - Thu, 15 Jan 2026 13:52:04 +0000 + EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience + https://arxiv.org/abs/2601.15876 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15876.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Taofeng Xue, Chong Peng, Mianqiu Huang, Linsen Guo, Tiancheng Han, Haozhe Wang, Jianing Wang, Xiaocheng Zhang, Xin Yang, Dengchang Zhao, Jinrui Ding, Xiandi Ma, Yuchen Xie, Peng Pei, Xunliang Cai, Xipeng Qiu</p><p><b>Upvotes:</b> 77</p><p><b>Summary:</b> The development of native computer-use agents (CUA) represents a significant leap in multimodal AI. However, their potential is currently bottlenecked by the constraints of static data scaling. Existing paradigms relying primarily on passive imitation of static datasets struggle to capture the intricate causal dynamics inherent in long-horizon computer tasks. In this work, we introduce EvoCUA, a native computer use agentic model. Unlike static imitation, EvoCUA integrates data generation and policy optimization into a self-sustaining evolutionary cycle. To mitigate data scarcity, we develop a verifiable synthesis engine that autonomously generates diverse tasks coupled with executable validators. To enable large-scale experience acquisition, we design a scalable infrastructure orchestrating tens of thousands of asynchronous sandbox rollouts. Building on these massive trajectories, we propose an iterative evolving learning strategy to efficiently internalize this experience. This mechanism dynamically regulates policy updates by identifying capability boundaries -- reinforcing successful routines while transforming failure trajectories into rich supervision through error analysis and self-correction. Empirical evaluations on the OSWorld benchmark demonstrate that EvoCUA achieves a success rate of 56.7%, establishing a new open-source state-of-the-art. Notably, EvoCUA significantly outperforms the previous best open-source model, OpenCUA-72B (45.0%), and surpasses leading closed-weights models such as UI-TARS-2 (53.1%). Crucially, our results underscore the generalizability of this approach: the evolving paradigm driven by learning from experience yields consistent performance gains across foundation models of varying scales, establishing a robust and scalable path for advancing native agent capabilities.</p> + https://arxiv.org/abs/2601.15876 + Thu, 22 Jan 2026 11:36:43 +0000 - Urban Socio-Semantic Segmentation with Vision-Language Reasoning - https://arxiv.org/abs/2601.10477 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10477.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yu Wang, Yi Wang, Rui Dai, Yujie Wang, Kaikui Liu, Xiangxiang Chu, Yansheng Li</p><p><b>Upvotes:</b> 148</p><p><b>Summary:</b> As hubs of human activity, urban surfaces consist of a wealth of semantic entities. Segmenting these various entities from satellite imagery is crucial for a range of downstream applications. Current advanced segmentation models can reliably segment entities defined by physical attributes (e.g., buildings, water bodies) but still struggle with socially defined categories (e.g., schools, parks). In this work, we achieve socio-semantic segmentation by vision-language model reasoning. To facilitate this, we introduce the Urban Socio-Semantic Segmentation dataset named SocioSeg, a new resource comprising satellite imagery, digital maps, and pixel-level labels of social semantic entities organized in a hierarchical structure. Additionally, we propose a novel vision-language reasoning framework called SocioReasoner that simulates the human process of identifying and annotating social semantic entities via cross-modal recognition and multi-stage reasoning. We employ reinforcement learning to optimize this non-differentiable process and elicit the reasoning capabilities of the vision-language model. Experiments demonstrate our approach's gains over state-of-the-art models and strong zero-shot generalization. Our dataset and code are available in https://github.com/AMAP-ML/SocioReasoner.</p> - https://arxiv.org/abs/2601.10477 - Thu, 15 Jan 2026 15:00:36 +0000 + Stable-DiffCoder: Pushing the Frontier of Code Diffusion Large Language Model + https://arxiv.org/abs/2601.15892 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.15892.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Chenghao Fan, Wen Heng, Bo Li, Sichen Liu, Yuxuan Song, Jing Su, Xiaoye Qu, Kai Shen, Wei Wei</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> Diffusion-based language models (DLLMs) offer non-sequential, block-wise generation and richer data reuse compared to autoregressive (AR) models, but existing code DLLMs still lag behind strong AR baselines under comparable budgets. We revisit this setting in a controlled study and introduce Stable-DiffCoder, a block diffusion code model that reuses the Seed-Coder architecture, data, and training pipeline. To enable efficient knowledge learning and stable training, we incorporate a block diffusion continual pretraining (CPT) stage enhanced by a tailored warmup and block-wise clipped noise schedule. Under the same data and architecture, Stable-DiffCoder overall outperforms its AR counterpart on a broad suite of code benchmarks. Moreover, relying only on the CPT and supervised fine-tuning stages, Stable-DiffCoder achieves stronger performance than a wide range of \~8B ARs and DLLMs, demonstrating that diffusion-based training can improve code modeling quality beyond AR training alone. Moreover, diffusion-based any-order modeling improves structured code modeling for editing and reasoning, and through data augmentation, benefits low-resource coding languages.</p> + https://arxiv.org/abs/2601.15892 + Thu, 22 Jan 2026 12:13:17 +0000 - A Safety Report on GPT-5.2, Gemini 3 Pro, Qwen3-VL, Doubao 1.8, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5 - https://arxiv.org/abs/2601.10527 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10527.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Xingjun Ma, Yixu Wang, Hengyuan Xu, Yutao Wu, Yifan Ding, Yunhan Zhao, Zilong Wang, Jiabin Hua, Ming Wen, Jianan Liu, Ranjie Duan, Yifeng Gao, Yingshui Tan, Yunhao Chen, Hui Xue, Xin Wang, Wei Cheng, Jingjing Chen, Zuxuan Wu, Bo Li, Yu-Gang Jiang</p><p><b>Upvotes:</b> 18</p><p><b>Summary:</b> The rapid evolution of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has produced substantial gains in reasoning, perception, and generative capability across language and vision. However, whether these advances yield commensurate improvements in safety remains unclear, in part due to fragmented evaluation practices limited to single modalities or threat models. In this report, we present an integrated safety evaluation of 7 frontier models: GPT-5.2, Gemini 3 Pro, Qwen3-VL, Doubao 1.8, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5. We evaluate each model across language, vision-language, and image generation settings using a unified protocol that integrates benchmark evaluation, adversarial evaluation, multilingual evaluation, and compliance evaluation. Aggregating our evaluations into safety leaderboards and model safety profiles across multiple evaluation modes reveals a sharply heterogeneous safety landscape. While GPT-5.2 demonstrates consistently strong and balanced safety performance across evaluations, other models exhibit pronounced trade-offs among benchmark safety, adversarial alignment, multilingual generalization, and regulatory compliance. Both language and vision-language modalities show significant vulnerability under adversarial evaluation, with all models degrading substantially despite strong results on standard benchmarks. Text-to-image models achieve relatively stronger alignment in regulated visual risk categories, yet remain brittle under adversarial or semantically ambiguous prompts. Overall, these results show that safety in frontier models is inherently multidimensional--shaped by modality, language, and evaluation scheme, underscoring the need for standardized safety evaluations to accurately assess real-world risk and guide responsible model development and deployment.</p> - https://arxiv.org/abs/2601.10527 - Thu, 15 Jan 2026 15:52:52 +0000 + SAMTok: Representing Any Mask with Two Words + https://arxiv.org/abs/2601.16093 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16093.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Yikang Zhou, Tao Zhang, Dengxian Gong, Yuanzheng Wu, Ye Tian, Haochen Wang, Haobo Yuan, Jiacong Wang, Lu Qi, Hao Fei, Anran Wang, Zhuochen Wang, Yujing Wang, Cheng Chen, Shunping Ji, Xiangtai Li</p><p><b>Upvotes:</b> 37</p><p><b>Summary:</b> Pixel-wise capabilities are essential for building interactive intelligent systems. However, pixel-wise multi-modal LLMs (MLLMs) remain difficult to scale due to complex region-level encoders, specialized segmentation decoders, and incompatible training objectives. To address these challenges, we present SAMTok, a discrete mask tokenizer that converts any region mask into two special tokens and reconstructs the mask using these tokens with high fidelity. By treating masks as new language tokens, SAMTok enables base MLLMs (such as the QwenVL series) to learn pixel-wise capabilities through standard next-token prediction and simple reinforcement learning, without architectural modifications and specialized loss design. SAMTok builds on SAM2 and is trained on 209M diverse masks using a mask encoder and residual vector quantizer to produce discrete, compact, and information-rich tokens. With 5M SAMTok-formatted mask understanding and generation data samples, QwenVL-SAMTok attains state-of-the-art or comparable results on region captioning, region VQA, grounded conversation, referring segmentation, scene graph parsing, and multi-round interactive segmentation. We further introduce a textual answer-matching reward that enables efficient reinforcement learning for mask generation, delivering substantial improvements on GRES and GCG benchmarks. Our results demonstrate a scalable and straightforward paradigm for equipping MLLMs with strong pixel-wise capabilities. Our code and models are available.</p> + https://arxiv.org/abs/2601.16093 + Thu, 22 Jan 2026 16:44:09 +0000 - HeartMuLa: A Family of Open Sourced Music Foundation Models - https://arxiv.org/abs/2601.10547 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10547.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Dongchao Yang, Yuxin Xie, Yuguo Yin, Zheyu Wang, Xiaoyu Yi, Gongxi Zhu, Xiaolong Weng, Zihan Xiong, Yingzhe Ma, Dading Cong, Jingliang Liu, Zihang Huang, Jinghan Ru, Rongjie Huang, Haoran Wan, Peixu Wang, Kuoxi Yu, Helin Wang, Liming Liang, Xianwei Zhuang, Yuanyuan Wang, Haohan Guo, Junjie Cao, Zeqian Ju, Songxiang Liu, Yuewen Cao, Heming Weng, Yuexian Zou</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> We present a family of open-source Music Foundation Models designed to advance large-scale music understanding and generation across diverse tasks and modalities. Our framework consists of four major components: (1) HeartCLAP, an audio-text alignment model; (2) HeartTranscriptor, a robust lyric recognition model optimized for real-world music scenarios; and (3) HeartCodec, a low-frame-rate (12.5 Hz) yet high-fidelity music codec tokenizer that captures long-range musical structure while preserving fine-grained acoustic details and enabling efficient autoregressive modeling; (4) HeartMuLa, an LLM-based song generation model capable of synthesizing high-fidelity music under rich, user-controllable conditions (e.g., textual style descriptions, lyrics, and reference audio). In addition, it provides two specialized modes: (i) fine-grained musical attribute control, which allows users to specify the style of different song sections (e.g., intro, verse, chorus) using natural language prompts; and (ii) short, engaging music generation, which is suitable as background music for short videos. Lastly, HeartMuLa improves significantly when scaled to 7B parameters. For the first time, we show that a Suno-level, commercial-grade system can be reproduced using academic-scale data and GPU resources. We expect these foundation models to serve as strong baselines for future research and to facilitate practical applications in multimodal content production.</p> - https://arxiv.org/abs/2601.10547 - Thu, 15 Jan 2026 16:14:25 +0000 + Rethinking Composed Image Retrieval Evaluation: A Fine-Grained Benchmark from Image Editing + https://arxiv.org/abs/2601.16125 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16125.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tingyu Song, Yanzhao Zhang, Mingxin Li, Zhuoning Guo, Dingkun Long, Pengjun Xie, Siyue Zhang, Yilun Zhao, Shu Wu</p><p><b>Upvotes:</b> 13</p><p><b>Summary:</b> Composed Image Retrieval (CIR) is a pivotal and complex task in multimodal understanding. Current CIR benchmarks typically feature limited query categories and fail to capture the diverse requirements of real-world scenarios. To bridge this evaluation gap, we leverage image editing to achieve precise control over modification types and content, enabling a pipeline for synthesizing queries across a broad spectrum of categories. Using this pipeline, we construct EDIR, a novel fine-grained CIR benchmark. EDIR encompasses 5,000 high-quality queries structured across five main categories and fifteen subcategories. Our comprehensive evaluation of 13 multimodal embedding models reveals a significant capability gap; even state-of-the-art models (e.g., RzenEmbed and GME) struggle to perform consistently across all subcategories, highlighting the rigorous nature of our benchmark. Through comparative analysis, we further uncover inherent limitations in existing benchmarks, such as modality biases and insufficient categorical coverage. Furthermore, an in-domain training experiment demonstrates the feasibility of our benchmark. This experiment clarifies the task challenges by distinguishing between categories that are solvable with targeted data and those that expose intrinsic limitations of current model architectures.</p> + https://arxiv.org/abs/2601.16125 + Thu, 22 Jan 2026 17:26:52 +0000 - Inference-time Physics Alignment of Video Generative Models with Latent World Models - https://arxiv.org/abs/2601.10553 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10553.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Jianhao Yuan, Xiaofeng Zhang, Felix Friedrich, Nicolas Beltran-Velez, Melissa Hall, Reyhane Askari-Hemmat, Xiaochuang Han, Nicolas Ballas, Michal Drozdzal, Adriana Romero-Soriano</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> State-of-the-art video generative models produce promising visual content yet often violate basic physics principles, limiting their utility. While some attribute this deficiency to insufficient physics understanding from pre-training, we find that the shortfall in physics plausibility also stems from suboptimal inference strategies. We therefore introduce WMReward and treat improving physics plausibility of video generation as an inference-time alignment problem. In particular, we leverage the strong physics prior of a latent world model (here, VJEPA-2) as a reward to search and steer multiple candidate denoising trajectories, enabling scaling test-time compute for better generation performance. Empirically, our approach substantially improves physics plausibility across image-conditioned, multiframe-conditioned, and text-conditioned generation settings, with validation from human preference study. Notably, in the ICCV 2025 Perception Test PhysicsIQ Challenge, we achieve a final score of 62.64%, winning first place and outperforming the previous state of the art by 7.42%. Our work demonstrates the viability of using latent world models to improve physics plausibility of video generation, beyond this specific instantiation or parameterization.</p> - https://arxiv.org/abs/2601.10553 - Thu, 15 Jan 2026 16:18:00 +0000 + ActionMesh: Animated 3D Mesh Generation with Temporal 3D Diffusion + https://arxiv.org/abs/2601.16148 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16148.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Remy Sabathier, David Novotny, Niloy J. Mitra, Tom Monnier</p><p><b>Upvotes:</b> 12</p><p><b>Summary:</b> Generating animated 3D objects is at the heart of many applications, yet most advanced works are typically difficult to apply in practice because of their limited setup, their long runtime, or their limited quality. We introduce ActionMesh, a generative model that predicts production-ready 3D meshes "in action" in a feed-forward manner. Drawing inspiration from early video models, our key insight is to modify existing 3D diffusion models to include a temporal axis, resulting in a framework we dubbed "temporal 3D diffusion". Specifically, we first adapt the 3D diffusion stage to generate a sequence of synchronized latents representing time-varying and independent 3D shapes. Second, we design a temporal 3D autoencoder that translates a sequence of independent shapes into the corresponding deformations of a pre-defined reference shape, allowing us to build an animation. Combining these two components, ActionMesh generates animated 3D meshes from different inputs like a monocular video, a text description, or even a 3D mesh with a text prompt describing its animation. Besides, compared to previous approaches, our method is fast and produces results that are rig-free and topology consistent, hence enabling rapid iteration and seamless applications like texturing and retargeting. We evaluate our model on standard video-to-4D benchmarks (Consistent4D, Objaverse) and report state-of-the-art performances on both geometric accuracy and temporal consistency, demonstrating that our model can deliver animated 3D meshes with unprecedented speed and quality.</p> + https://arxiv.org/abs/2601.16148 + Thu, 22 Jan 2026 17:41:13 +0000 - Action100M: A Large-scale Video Action Dataset - https://arxiv.org/abs/2601.10592 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10592.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Delong Chen, Tejaswi Kasarla, Yejin Bang, Mustafa Shukor, Willy Chung, Jade Yu, Allen Bolourchi, Theo Moutakanni, Pascale Fung</p><p><b>Upvotes:</b> 15</p><p><b>Summary:</b> Inferring physical actions from visual observations is a fundamental capability for advancing machine intelligence in the physical world. Achieving this requires large-scale, open-vocabulary video action datasets that span broad domains. We introduce Action100M, a large-scale dataset constructed from 1.2M Internet instructional videos (14.6 years of duration), yielding O(100 million) temporally localized segments with open-vocabulary action supervision and rich captions. Action100M is generated by a fully automated pipeline that (i) performs hierarchical temporal segmentation using V-JEPA 2 embeddings, (ii) produces multi-level frame and segment captions organized as a Tree-of-Captions, and (iii) aggregates evidence with a reasoning model (GPT-OSS-120B) under a multi-round Self-Refine procedure to output structured annotations (brief/detailed action, actor, brief/detailed caption). Training VL-JEPA on Action100M demonstrates consistent data-scaling improvements and strong zero-shot performance across diverse action recognition benchmarks, establishing Action100M as a new foundation for scalable research in video understanding and world modeling.</p> - https://arxiv.org/abs/2601.10592 - Thu, 15 Jan 2026 17:02:27 +0000 + Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning + https://arxiv.org/abs/2601.16163 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16163.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Moo Jin Kim, Yihuai Gao, Tsung-Yi Lin, Yen-Chen Lin, Yunhao Ge, Grace Lam, Percy Liang, Shuran Song, Ming-Yu Liu, Chelsea Finn, Jinwei Gu</p><p><b>Upvotes:</b> 11</p><p><b>Summary:</b> Recent video generation models demonstrate remarkable ability to capture complex physical interactions and scene evolution over time. To leverage their spatiotemporal priors, robotics works have adapted video models for policy learning but introduce complexity by requiring multiple stages of post-training and new architectural components for action generation. In this work, we introduce Cosmos Policy, a simple approach for adapting a large pretrained video model (Cosmos-Predict2) into an effective robot policy through a single stage of post-training on the robot demonstration data collected on the target platform, with no architectural modifications. Cosmos Policy learns to directly generate robot actions encoded as latent frames within the video model's latent diffusion process, harnessing the model's pretrained priors and core learning algorithm to capture complex action distributions. Additionally, Cosmos Policy generates future state images and values (expected cumulative rewards), which are similarly encoded as latent frames, enabling test-time planning of action trajectories with higher likelihood of success. In our evaluations, Cosmos Policy achieves state-of-the-art performance on the LIBERO and RoboCasa simulation benchmarks (98.5% and 67.1% average success rates, respectively) and the highest average score in challenging real-world bimanual manipulation tasks, outperforming strong diffusion policies trained from scratch, video model-based policies, and state-of-the-art vision-language-action models fine-tuned on the same robot demonstrations. Furthermore, given policy rollout data, Cosmos Policy can learn from experience to refine its world model and value function and leverage model-based planning to achieve even higher success rates in challenging tasks. We release code, models, and training data at https://research.nvidia.com/labs/dir/cosmos-policy/</p> + https://arxiv.org/abs/2601.16163 + Thu, 22 Jan 2026 18:09:30 +0000 - Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding - https://arxiv.org/abs/2601.10611 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10611.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Christopher Clark, Jieyu Zhang, Zixian Ma, Jae Sung Park, Mohammadreza Salehi, Rohun Tripathi, Sangho Lee, Zhongzheng Ren, Chris Dongjoo Kim, Yinuo Yang, Vincent Shao, Yue Yang, Weikai Huang, Ziqi Gao, Taira Anderson, Jianrui Zhang, Jitesh Jain, George Stoica, Winson Han, Ali Farhadi, Ranjay Krishna</p><p><b>Upvotes:</b> 20</p><p><b>Summary:</b> Today's strongest video-language models (VLMs) remain proprietary. The strongest open-weight models either rely on synthetic data from proprietary VLMs, effectively distilling from them, or do not disclose their training data or recipe. As a result, the open-source community lacks the foundations needed to improve on the state-of-the-art video (and image) language models. Crucially, many downstream applications require more than just high-level video understanding; they require grounding -- either by pointing or by tracking in pixels. Even proprietary models lack this capability. We present Molmo2, a new family of VLMs that are state-of-the-art among open-source models and demonstrate exceptional new capabilities in point-driven grounding in single image, multi-image, and video tasks. Our key contribution is a collection of 7 new video datasets and 2 multi-image datasets, including a dataset of highly detailed video captions for pre-training, a free-form video Q&A dataset for fine-tuning, a new object tracking dataset with complex queries, and an innovative new video pointing dataset, all collected without the use of closed VLMs. We also present a training recipe for this data utilizing an efficient packing and message-tree encoding scheme, and show bi-directional attention on vision tokens and a novel token-weight strategy improves performance. Our best-in-class 8B model outperforms others in the class of open weight and data models on short videos, counting, and captioning, and is competitive on long-videos. On video-grounding Molmo2 significantly outperforms existing open-weight models like Qwen3-VL (35.5 vs 29.6 accuracy on video counting) and surpasses proprietary models like Gemini 3 Pro on some tasks (38.4 vs 20.0 F1 on video pointing and 56.2 vs 41.1 J&F on video tracking).</p> - https://arxiv.org/abs/2601.10611 - Thu, 15 Jan 2026 17:27:44 +0000 + Learning to Discover at Test Time + https://arxiv.org/abs/2601.16175 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16175.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Mert Yuksekgonul, Daniel Koceja, Xinhao Li, Federico Bianchi, Jed McCaleb, Xiaolong Wang, Jan Kautz, Yejin Choi, James Zou, Carlos Guestrin, Yu Sun</p><p><b>Upvotes:</b> 35</p><p><b>Summary:</b> How can we use AI to discover a new state of the art for a scientific problem? Prior work in test-time scaling, such as AlphaEvolve, performs search by prompting a frozen LLM. We perform reinforcement learning at test time, so the LLM can continue to train, but now with experience specific to the test problem. This form of continual learning is quite special, because its goal is to produce one great solution rather than many good ones on average, and to solve this very problem rather than generalize to other problems. Therefore, our learning objective and search subroutine are designed to prioritize the most promising solutions. We call this method Test-Time Training to Discover (TTT-Discover). Following prior work, we focus on problems with continuous rewards. We report results for every problem we attempted, across mathematics, GPU kernel engineering, algorithm design, and biology. TTT-Discover sets the new state of the art in almost all of them: (i) Erdős' minimum overlap problem and an autocorrelation inequality; (ii) a GPUMode kernel competition (up to 2times faster than prior art); (iii) past AtCoder algorithm competitions; and (iv) denoising problem in single-cell analysis. Our solutions are reviewed by experts or the organizers. All our results are achieved with an open model, OpenAI gpt-oss-120b, and can be reproduced with our publicly available code, in contrast to previous best results that required closed frontier models. Our test-time training runs are performed using Tinker, an API by Thinking Machines, with a cost of only a few hundred dollars per problem.</p> + https://arxiv.org/abs/2601.16175 + Thu, 22 Jan 2026 18:24:00 +0000 - PACEvolve: Enabling Long-Horizon Progress-Aware Consistent Evolution - https://arxiv.org/abs/2601.10657 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10657.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Minghao Yan, Bo Peng, Benjamin Coleman, Ziqi Chen, Zhouhang Xie, Zhankui He, Noveen Sachdeva, Isabella Ye, Weili Wang, Chi Wang, Ed H. Chi, Wang-Cheng Kang, Derek Zhiyuan Cheng, Beidou Wang</p><p><b>Upvotes:</b> 17</p><p><b>Summary:</b> Large Language Models (LLMs) have emerged as powerful operators for evolutionary search, yet the design of efficient search scaffolds remains ad hoc. While promising, current LLM-in-the-loop systems lack a systematic approach to managing the evolutionary process. We identify three distinct failure modes: Context Pollution, where experiment history biases future candidate generation; Mode Collapse, where agents stagnate in local minima due to poor exploration-exploitation balance; and Weak Collaboration, where rigid crossover strategies fail to leverage parallel search trajectories effectively. We introduce Progress-Aware Consistent Evolution (PACEvolve), a framework designed to robustly govern the agent's context and search dynamics, to address these challenges. PACEvolve combines hierarchical context management (HCM) with pruning to address context pollution; momentum-based backtracking (MBB) to escape local minima; and a self-adaptive sampling policy that unifies backtracking and crossover for dynamic search coordination (CE), allowing agents to balance internal refinement with cross-trajectory collaboration. We demonstrate that PACEvolve provides a systematic path to consistent, long-horizon self-improvement, achieving state-of-the-art results on LLM-SR and KernelBench, while discovering solutions surpassing the record on Modded NanoGPT.</p> - https://arxiv.org/abs/2601.10657 - Thu, 15 Jan 2026 18:25:23 +0000 + 360Anything: Geometry-Free Lifting of Images and Videos to 360° + https://arxiv.org/abs/2601.16192 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16192.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Ziyi Wu, Daniel Watson, Andrea Tagliasacchi, David J. Fleet, Marcus A. Brubaker, Saurabh Saxena</p><p><b>Upvotes:</b> 7</p><p><b>Summary:</b> Lifting perspective images and videos to 360° panoramas enables immersive 3D world generation. Existing approaches often rely on explicit geometric alignment between the perspective and the equirectangular projection (ERP) space. Yet, this requires known camera metadata, obscuring the application to in-the-wild data where such calibration is typically absent or noisy. We propose 360Anything, a geometry-free framework built upon pre-trained diffusion transformers. By treating the perspective input and the panorama target simply as token sequences, 360Anything learns the perspective-to-equirectangular mapping in a purely data-driven way, eliminating the need for camera information. Our approach achieves state-of-the-art performance on both image and video perspective-to-360° generation, outperforming prior works that use ground-truth camera information. We also trace the root cause of the seam artifacts at ERP boundaries to zero-padding in the VAE encoder, and introduce Circular Latent Encoding to facilitate seamless generation. Finally, we show competitive results in zero-shot camera FoV and orientation estimation benchmarks, demonstrating 360Anything's deep geometric understanding and broader utility in computer vision tasks. Additional results are available at https://360anything.github.io/.</p> + https://arxiv.org/abs/2601.16192 + Thu, 22 Jan 2026 18:45:59 +0000 - MatchTIR: Fine-Grained Supervision for Tool-Integrated Reasoning via Bipartite Matching - https://arxiv.org/abs/2601.10712 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10712.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Changle Qu, Sunhao Dai, Hengyi Cai, Jun Xu, Shuaiqiang Wang, Dawei Yin</p><p><b>Upvotes:</b> 19</p><p><b>Summary:</b> Tool-Integrated Reasoning (TIR) empowers large language models (LLMs) to tackle complex tasks by interleaving reasoning steps with external tool interactions. However, existing reinforcement learning methods typically rely on outcome- or trajectory-level rewards, assigning uniform advantages to all steps within a trajectory. This coarse-grained credit assignment fails to distinguish effective tool calls from redundant or erroneous ones, particularly in long-horizon multi-turn scenarios. To address this, we propose MatchTIR, a framework that introduces fine-grained supervision via bipartite matching-based turn-level reward assignment and dual-level advantage estimation. Specifically, we formulate credit assignment as a bipartite matching problem between predicted and ground-truth traces, utilizing two assignment strategies to derive dense turn-level rewards. Furthermore, to balance local step precision with global task success, we introduce a dual-level advantage estimation scheme that integrates turn-level and trajectory-level signals, assigning distinct advantage values to individual interaction turns. Extensive experiments on three benchmarks demonstrate the superiority of MatchTIR. Notably, our 4B model surpasses the majority of 8B competitors, particularly in long-horizon and multi-turn tasks. Our codes are available at https://github.com/quchangle1/MatchTIR.</p> - https://arxiv.org/abs/2601.10712 - Thu, 15 Jan 2026 18:59:23 +0000 + LLM-in-Sandbox Elicits General Agentic Intelligence + https://arxiv.org/abs/2601.16206 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16206.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Daixuan Cheng, Shaohan Huang, Yuxian Gu, Huatong Song, Guoxin Chen, Li Dong, Wayne Xin Zhao, Ji-Rong Wen, Furu Wei</p><p><b>Upvotes:</b> 64</p><p><b>Summary:</b> We introduce LLM-in-Sandbox, enabling LLMs to explore within a code sandbox (i.e., a virtual computer), to elicit general intelligence in non-code domains. We first demonstrate that strong LLMs, without additional training, exhibit generalization capabilities to leverage the code sandbox for non-code tasks. For example, LLMs spontaneously access external resources to acquire new knowledge, leverage the file system to handle long contexts, and execute scripts to satisfy formatting requirements. We further show that these agentic capabilities can be enhanced through LLM-in-Sandbox Reinforcement Learning (LLM-in-Sandbox-RL), which uses only non-agentic data to train models for sandbox exploration. Experiments demonstrate that LLM-in-Sandbox, in both training-free and post-trained settings, achieves robust generalization spanning mathematics, physics, chemistry, biomedicine, long-context understanding, and instruction following. Finally, we analyze LLM-in-Sandbox's efficiency from computational and system perspectives, and open-source it as a Python package to facilitate real-world deployment.</p> + https://arxiv.org/abs/2601.16206 + Thu, 22 Jan 2026 18:57:09 +0000 - Alterbute: Editing Intrinsic Attributes of Objects in Images - https://arxiv.org/abs/2601.10714 - <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.10714.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Tal Reiss, Daniel Winter, Matan Cohen, Alex Rav-Acha, Yael Pritch, Ariel Shamir, Yedid Hoshen</p><p><b>Upvotes:</b> 21</p><p><b>Summary:</b> We introduce Alterbute, a diffusion-based method for editing an object's intrinsic attributes in an image. We allow changing color, texture, material, and even the shape of an object, while preserving its perceived identity and scene context. Existing approaches either rely on unsupervised priors that often fail to preserve identity or use overly restrictive supervision that prevents meaningful intrinsic variations. Our method relies on: (i) a relaxed training objective that allows the model to change both intrinsic and extrinsic attributes conditioned on an identity reference image, a textual prompt describing the target intrinsic attributes, and a background image and object mask defining the extrinsic context. At inference, we restrict extrinsic changes by reusing the original background and object mask, thereby ensuring that only the desired intrinsic attributes are altered; (ii) Visual Named Entities (VNEs) - fine-grained visual identity categories (e.g., ''Porsche 911 Carrera'') that group objects sharing identity-defining features while allowing variation in intrinsic attributes. We use a vision-language model to automatically extract VNE labels and intrinsic attribute descriptions from a large public image dataset, enabling scalable, identity-preserving supervision. Alterbute outperforms existing methods on identity-preserving object intrinsic attribute editing.</p> - https://arxiv.org/abs/2601.10714 - Thu, 15 Jan 2026 18:59:53 +0000 + Scaling Text-to-Image Diffusion Transformers with Representation Autoencoders + https://arxiv.org/abs/2601.16208 + <p><img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2601.16208.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /></p><p><b>Authors:</b> Shengbang Tong, Boyang Zheng, Ziteng Wang, Bingda Tang, Nanye Ma, Ellis Brown, Jihan Yang, Rob Fergus, Yann LeCun, Saining Xie</p><p><b>Upvotes:</b> 46</p><p><b>Summary:</b> Representation Autoencoders (RAEs) have shown distinct advantages in diffusion modeling on ImageNet by training in high-dimensional semantic latent spaces. In this work, we investigate whether this framework can scale to large-scale, freeform text-to-image (T2I) generation. We first scale RAE decoders on the frozen representation encoder (SigLIP-2) beyond ImageNet by training on web, synthetic, and text-rendering data, finding that while scale improves general fidelity, targeted data composition is essential for specific domains like text. We then rigorously stress-test the RAE design choices originally proposed for ImageNet. Our analysis reveals that scaling simplifies the framework: while dimension-dependent noise scheduling remains critical, architectural complexities such as wide diffusion heads and noise-augmented decoding offer negligible benefits at scale Building on this simplified framework, we conduct a controlled comparison of RAE against the state-of-the-art FLUX VAE across diffusion transformer scales from 0.5B to 9.8B parameters. RAEs consistently outperform VAEs during pretraining across all model scales. Further, during finetuning on high-quality datasets, VAE-based models catastrophically overfit after 64 epochs, while RAE models remain stable through 256 epochs and achieve consistently better performance. Across all experiments, RAE-based diffusion models demonstrate faster convergence and better generation quality, establishing RAEs as a simpler and stronger foundation than VAEs for large-scale T2I generation. Additionally, because both visual understanding and generation can operate in a shared representation space, the multimodal model can directly reason over generated latents, opening new possibilities for unified models.</p> + https://arxiv.org/abs/2601.16208 + Thu, 22 Jan 2026 18:58:16 +0000