clean: remove Hermes cron artifacts, simplify for EasyPanel deploy

.dockerignore

@@ -0,0 +1,8 @@
+.venv/
+__pycache__/
+*.py[cod]
+*.egg-info/
+.git/
+.github/
+test_*/
+.DS_Store

.gitignore

@@ -1,37 +1,5 @@
-# Byte-compiled / optimized files
+# Python
 __pycache__/
 *.py[cod]
-*$py.class
-
-# Distribution / packaging
-*.egg-info/
-dist/
-build/
-
-# Virtual environments
-venv/
-env/
-.env/
 .venv/
-
+venv/
-# IDE specific files
-.idea/
-.vscode/
-*.swp
-*.swo
-
-# OS specific files
-.DS_Store
-Thumbs.db
-
-# Local testing/development
-*.bak
-*.tmp
-.coverage
-htmlcov/
-
-# Exclude the HTML file downloaded from HF (if you keep test files locally)
-*Daily Papers - Hugging Face.html
-
-# Uncomment the line below if you don't want to commit generated feed files
-# *.xml

Dockerfile

@@ -0,0 +1,38 @@
+# ─────────────────────────────────────────────
+# HuggingFace Daily Paper RSS — EasyPanel-ready Dockerfile
+# ─────────────────────────────────────────────
+# Single container that:
+#   1. Generates RSS feeds on startup
+#   2. Serves them via HTTP (port 8080, CORS enabled)
+#   3. Auto-refreshes every 24h (background thread)
+#
+# Deploy on EasyPanel: health check + persistent volume
+# ─────────────────────────────────────────────
+
+FROM python:3.11-slim
+
+WORKDIR /app
+
+# System deps
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    ca-certificates \
+    && rm -rf /var/lib/apt/lists/*
+
+# Python deps
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# App code
+COPY parser.py rss_generator.py run.py server.py ./
+
+# Volume for persistent feed output
+VOLUME ["/data"]
+
+# Health check — verify feed.xml exists (generated on startup)
+HEALTHCHECK --interval=60s --timeout=10s --start-period=30s --retries=3 \
+    CMD python -c "import os; exit(0 if os.path.exists('/data/feed.xml') else 1)"
+
+EXPOSE 8080
+
+ENTRYPOINT ["python", "server.py"]
+CMD ["--dir", "/data", "--port", "8080", "--refresh", "24"]

feed.xml

@@ -7,244 +7,181 @@
     <docs>http://www.rssboard.org/rss-specification</docs>
     <generator>python-feedgen</generator>
     <language>en</language>
-    <lastBuildDate>Mon, 22 Jun 2026 00:24:34 +0000</lastBuildDate>
+    <lastBuildDate>Sat, 27 Jun 2026 07:38:14 +0000</lastBuildDate>
     <item>
-      <title>Duration Aware Scheduling for ASR Serving Under Workload Drift</title>
+      <title>How Post-Training Shapes Biological Reasoning Models</title>
-      <link>https://arxiv.org/abs/2603.11273</link>
+      <link>https://arxiv.org/abs/2606.16517</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2603.11273.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Darshan Makwana, Yash Jogi, Harsh Kotta, Aayush Kubba&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Scheduling policies in large-scale Automatic Speech Recognition (ASR) serving pipelines play a key role in determining end-to-end (E2E) latency. Yet, widely used serving engines rely on first-come-first-served (FCFS) scheduling, which ignores variability in request duration and leads to head-of-line blocking under workload drift. We show that audio duration is an accurate proxy for job processing time in ASR models such as Whisper, and use this insight to enable duration-aware scheduling. We integrate two classical algorithms, Shortest Job First (SJF) and Highest Response Ratio Next (HRRN), into vLLM and evaluate them under realistic and drifted workloads. On LibriSpeech test-clean, compared to baseline, SJF reduces median E2E latency by up to 73% at high load, but increases 90th-percentile tail latency by up to 97% due to starvation of long requests. HRRN addresses this trade-off: it reduces median E2E latency by up to 28% while bounding tail-latency degradation to at most 24%. These gains persist under workload drift, with no throughput penalty and &lt;0.1\,ms scheduling overhead per request.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.16517.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Lukas Fesser, Hanlin Zhang, Michelle M. Li, Eric Wang, Bryan Perozzi, Shekoofeh Azizi, Sham M. Kakade, Marinka Zitnik&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Scientific reasoning models for biology combine language models with foundation models trained on multimodal biological data, including DNA, RNA, and proteins. These models are built through post-training, yet how each stage shapes reasoning and generalization remains poorly understood. We study when post-training improves performance and when it induces over-specialization. Across genomics, transcriptomics, and proteins, we train and evaluate more than 100 biological reasoning models under controlled variation in backbone, continued pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning (RL), measuring both in-domain (ID) and out-of-domain (OOD) performance. We find that each post-training stage reshapes generalization in a distinct way rather than contributing uniform gains. CPT improves downstream performance by aligning models with biological language. SFT consistently increases ID performance but causes OOD performance to peak early and decline as models fit the training distribution. RL, when applied to strong SFT checkpoints with aligned rewards, improves OOD performance and partially recovers generalization. These results show that biological reasoning does not improve monotonically with additional supervision or compute. Instead, performance depends on how training stages are composed. Under fixed post-training budgets, the strongest ID-OOD trade-off comes from brief SFT, larger RL allocations, and asymmetric adaptation capacity across stages.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2603.11273</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.16517</guid>
-      <pubDate>Wed, 11 Mar 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>ReSyn: A Generalized Recursive Regular Expression Synthesis Framework</title>
-      <link>https://arxiv.org/abs/2603.24624</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2603.24624.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Seongmin Kim, Hyunjoon Cheon, Su-Hyeon Kim, Yo-Sub Han, Sang-Ki Ko&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 2&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Existing Programming-By-Example (PBE) systems often rely on simplified benchmarks that fail to capture the high structural complexity of real-world regexes, such as deeper nesting and frequent use of union operations. To overcome the resulting performance drop, we propose ReSyn, a synthesizer-agnostic divide-and-conquer framework that decomposes complex synthesis problem into manageable sub-problems. We also introduce Set2Regex, a parameter-efficient synthesizer capturing the permutation invariance of examples. Experimental results demonstrate that ReSyn significantly boosts accuracy across various synthesizers, and its combination with Set2Regex establishes a new state-of-the-art on challenging real-world benchmark. The complete source code, datasets, and pre-trained model checkpoints are publicly available at https://github.com/mrseongminkim/ReSyn.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2603.24624</guid>
-      <pubDate>Sat, 13 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>DragMesh-2: Physically Plausible Dexterous Hand-Object Interaction with Articulated Objects</title>
-      <link>https://arxiv.org/abs/2606.15133</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.15133.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Tianshan Zhang, Yijia Duan, Yanjun Li, Zeyu Zhang, Hao Tang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 69&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Dexterous interaction with articulated objects is important for household, assistive, and humanoid manipulation, where multi-finger hands can provide compliant contact patterns beyond parallel-jaw grasping. However, articulated-object manipulation differs from static-object manipulation: the target part cannot be directly actuated, and its motion must emerge through sustained physical hand--handle contact. This makes the transition from object-centric articulated generation to hand-driven dexterous hand--object interaction non-trivial, since geometric trajectory replay or open-loop execution does not model the contact dynamics required to move the articulated part. Moreover, policies trained only for task completion under fixed dynamics can overfit nominal contact loads, especially without tactile or force feedback, and may degrade when the contact load changes. To address these challenges, we present DragMesh-2, a contact-driven framework for dexterous interaction with articulated objects that extends articulated interaction from object-centric generation to hand-driven dexterous hand--object interaction, where articulated motion must arise through physical contact. We further propose PICA, a physically informed contact-aware training mechanism that injects physical signals into policy learning without tactile or force feedback, improving robustness and task success under changing contact loads. Finally, we conduct systematic evaluation across multiple damping conditions and articulated-object categories to study robustness under contact-load variation, and provide a pure-geometry dexterous interaction resource to support future loco-manipulation and humanoid hand--object interaction research. Across seven GAPartNet objects, DragMesh-2 achieves stronger robustness under contact-load variation than the compared methods while maintaining high task success across damping conditions.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.15133</guid>
-      <pubDate>Sat, 13 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>Selective Synergistic Learning for Video Object-Centric Learning</title>
-      <link>https://arxiv.org/abs/2606.15527</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.15527.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; WonJun Moon, Jae-Pil Heo&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Typical video object-centric learning (VOCL) approaches employ slot-based frameworks that rely on reconstruction-driven encoder-decoder architectures, where learning is mediated by two spatial maps: attention maps from the encoder and object maps from the decoder. As these two distinct maps exhibit different properties, a recent dense alignment strategy attempted to reconcile this discrepancy by enforcing agreement across all spatio-temporal patches via contrastive learning. However, this indiscriminate alignment inadvertently propagates the inherent weaknesses of each module, such as noisy encoder predictions and blurred decoder boundaries. Moreover, computing dense similarities across all pairs incurs a computational cost quadratic in the total number of spatio-temporal patches, severely limiting scalability. Motivated by this, we propose Selective Synergistic Learning (SSync). Instead of exhaustive patch-to-patch alignment, SSync prevents error propagation by selectively distilling only the most reliable cues: leveraging the encoder strictly for boundary refinement and the decoder for interior denoising. This is realized via a pseudo-labeling with linear complexity, eliminating the need for quadratic spatial comparisons. Also, to prevent the reinforcement of architectural biases like slot redundancy, we introduce a transitive pseudo-label merging that consolidates overlapping slots based on spatio-temporal activation consistency. Extensive studies demonstrate that SSync improves decomposition quality and serves as a versatile, plug-and-play module while also exhibiting exceptional robustness to slot configurations. Code is available at github.com/wjun0830/SSync.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.15527</guid>
-      <pubDate>Sun, 14 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>The Data Manifold under the Microscope</title>
-      <link>https://arxiv.org/abs/2606.15760</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.15760.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Marios Koulakis, Constantin Seibold&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; A significant gap exists between theory and practice in deep learning. Generalization and approximation error bounds are often derived for simplified models or are too loose to be informative. Many rely on the manifold hypothesis and on geometric regularity such as intrinsic dimension, curvature, and reach. Progress requires insight into data-manifold geometry and suitable benchmarks, yet existing options are polarized: analytic manifolds with known geometry but limited applicability, or real-world datasets where geometry is only coarsely estimable. We introduce a benchmarking framework for studying data geometry. We repurpose and extend dSprites and COIL-20 with additional transformation dimensions and dense, axis-aligned sampling, and pair them with finite-difference estimators that recover curvature, reach, and volume at near-ground-truth accuracy in a regime where general-purpose estimators are unreliable or difficult to deploy. The framework is intended as a controlled testbed, useful as a calibration environment for geometric estimators and a sandbox for probing theoretical assumptions. To illustrate its use, we present two application studies, namely assessing the scaling behavior of the bounds of Genovese et al. and Fefferman et al., and tracking the layer-wise geometry of a β-VAE, highlighting the behavior of current bounds and the value of controlled benchmarks for guiding and validating future theory.   A reference implementation is available at https://github.com/koulakis/manifold-microscope.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.15760</guid>
-      <pubDate>Sun, 14 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>No Resource, No Benchmarks, No Problem? Evaluating and Improving LLMs for Code Generation in No-Resource Languages</title>
-      <link>https://arxiv.org/abs/2606.16827</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.16827.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Alessandro Giagnorio, Alberto Martin-Lopez, Gabriele Bavota&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Large Language Models (LLMs) have significantly advanced the automation of software engineering tasks. One prominent example is code generation, where an LLM produces code in a specified programming language based on a natural language description. Most research in this area has focused on high-resource languages, such as Python or Java, which benefit from abundant training data. A smaller body of work has explored low-resource languages, which are underrepresented in training corpora. In contrast, no-resource languages for which LLMs have seen virtually no training data remain largely unstudied. These languages often emerge in industry, where organizations develop proprietary or domain-specific languages unsupported by commercial tools like GitHub Copilot. This results in the need for companies to deploy their own in-house code recommenders. To investigate possible solutions in this context, we build and release three code generation benchmarks for no-resource languages, based on two recently proposed programming languages for which very little training data is available. Using these benchmarks, we experiment several solutions to teach LLMs about no-resource languages, including prompt-based techniques as well as pre-training and fine-tuning exploiting the little data available. While further pre-training gives the largest performance gains for no-resource languages, applying it directly to instruction-tuned models harms their ability to follow instructions. To address this, we start from a base model, further pre-training it on the target language, and then inject instruction-following capabilities via weight diff transfer from an instruction model. Such an approach significantly improves code generation capabilities in no-resource settings, allowing companies to cheaply deploy a specialized instruct model without dealing with the computational cost of instruction fine-tuning.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.16827</guid>
       <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Taylor-Calibrate: Principled Initialization for Hybrid Linear Attention Distillation</title>
+      <title>CoffeeBench: Benchmarking Long-Horizon LLM Agents in Heterogeneous Multi-Agent Economies</title>
-      <link>https://arxiv.org/abs/2606.16429</link>
+      <link>https://arxiv.org/abs/2606.16613</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.16429.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Zhongzhu Zhou, Qingyang Wu, Junxiong Wang, Mayank Mishra, Shuaiwen Leon Song, Ben Athiwaratkun, Chenfeng Xu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Hybrid linear attention models offer an appealing path to faster long-context inference: they reduce the quadratic cost and KV-cache burden of full softmax attention while retaining much of the quality of Transformer models. A practical way to obtain such models is to convert a pretrained Transformer instead of pretraining a new architecture from scratch, but this conversion is still brittle. Simply copying the teacher attention projections into a Gated DeltaNet (GDN) student does not specify the new recurrent decay, write, and output-gating dynamics. As a result, the converted model often starts in a poor dynamical regime and must spend many distillation tokens repairing initialization rather than learning the remaining teacher behavior. We propose Taylor-Calibrate, a lightweight initialization method for hybrid GDN students. The method uses Taylor-guided teacher attention statistics to set the value projection, memory timescale, write gates, and output gate, then applies a short per-layer alignment step to match each converted layer to the teacher output. Across four teacher settings and three retained-layer policies, Taylor-Calibrate gives substantially stronger zero-shot students, with up to an 88x improvement in a representative ablation, and reaches matched recovery targets with 4.9x--9.2x fewer training tokens than naive conversion.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.16613.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Issa Sugiura, Daichi Hattori, Kazuo Araragi, Keita Ogawa, Shota Onose, Taro Makino, Teppei Usuki, Takashi Ishida&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 7&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; As LLM agents become capable of increasingly long-horizon tasks, evaluating their performance in economic systems is becoming increasingly important. Unlike existing benchmarks that primarily evaluate a single agent interacting with a passive environment, economic systems are inherently multi-agent, requiring autonomous agents to communicate, negotiate, and transact while pursuing their own objectives over extended periods. We introduce CoffeeBench, a benchmark for evaluating LLM agents in a long-horizon multi-agent economy composed of heterogeneous firms. In CoffeeBench, two farmers, two roasters, and two retailers autonomously operate their businesses over a 90-day simulation, each seeking to maximize cumulative net income through communication and transactions while managing cash, inventory, and pricing. The evaluated model controls one coffee roaster, while the remaining firms are controlled by fixed reference agents. Across several recent open-weight and proprietary LLMs, all models outperform a passive baseline that takes no actions, with most achieving positive net income. Analysis of agent behavior reveals substantial differences in long-horizon economic interaction: higher-performing models communicate more actively with other firms, whereas Claude~Haiku~4.5 exhibits an idle-drift failure mode, repeatedly choosing inaction despite producing coherent assessments and plans. We release our code and agent trajectories to support future research.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.16429</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.16613</guid>
       <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Understanding the Behaviors of Environment-aware Information Retrieval</title>
+      <title>Discretizing Reward Models</title>
-      <link>https://arxiv.org/abs/2606.16817</link>
+      <link>https://arxiv.org/abs/2606.21795</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.16817.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Ruifeng Yuan, Chaohao Yuan, David Dai, Yu Rong, Hong Cheng, Hou Pong Chan, Chenghao Xiao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Recent retrieval-augmented generation (RAG) approaches have demonstrated strong capability in handling complex queries, yet current research overlooks a critical challenge: different retrievers require fundamentally different query formulation strategies for optimal performance. In this work, we present the first systematic analysis of how LLMs can learn to adapt their query formulation strategies for different retrievers via reinforcement learning (RL). Our empirical study reveals that RL effectively teaches an LLM to tailor its queries to specific retriever characteristics. We discover that different retrievers exhibit surprisingly distinct optimal query styles (e.g., descriptive vs. question-like), suggesting strategies learned for one retriever ineffective for another. We further show that performance can be enhanced by incorporating retriever-specific human guidance and by scaling model size. To facilitate learning over multi-retrieval-step trajectories, we introduce a branching-based rollout technique that improves training stability. Our work provides the first empirical evidence and actionable insights for building truly retriever-aware RAG systems. Code and resources are available at https://github.com/LCO-Embedding/Envs-aware-Information-Retrieval.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.21795.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Vijay Viswanathan, Shiqi Wang, Devamanyu Hazarika, Chirag Nagpal, Tongshuang Wu, Graham Neubig, Yuning Mao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Despite their widespread use, the role of reward models in shaping reinforcement learning is poorly understood. Reward models offer a tempting promise: they automatically estimate response quality in the absence of verifiers or human judges. Unlike "verifiable rewards" which typically produce binary scores, reward models typically produce continuous scores, allowing them to be sensitive to fine-grained differences in responses. However, we show this apparent strength is a serious weakness: many popular reward models are oversensitive, assigning different scores to equally good responses. Theoretically, we show that seemingly perfect reward models can be highly oversensitive; empirically, this oversensitivity can lead to bad policies. In place of existing notions of "reward model accuracy," we propose evaluating reward models using distinct measures of "discriminative ability" and "specificity" (the complement of oversensitivity). As a solution, we describe a training-free algorithm that uses Monte Carlo dropout on any neural reward model to produce discrete reward clusters. Theoretically, we prove there exist discretizations that reduce oversensitivity at minimal expense of discriminative ability; empirically we show, in both controlled and natural RL settings, that discretizing rewards leads to less reward hacking and better policies than training on the original rewards.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.16817</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.21795</guid>
-      <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Fri, 19 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Thinking with Visual Grounding</title>
+      <title>OpenBioRQ: Unsolved Biomedical Research Questions for Agents</title>
-      <link>https://arxiv.org/abs/2606.16122</link>
+      <link>https://arxiv.org/abs/2606.21959</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.16122.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Junkai Zhang, Yihe Deng, Kai-Wei Chang, Wei Wang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 10&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Visual thinking should not only sound right; it should show its evidence. While recent vision-language models (VLMs) can produce natural-language reasoning traces, these traces often leave the supporting image regions implicit, making them hard to verify and difficult to supervise. We introduce visually grounded thinking, a reasoning process in which models interleave natural-language thoughts with explicit point or box groundings of the visual evidence used at each step. This lets the model express intermediate reasoning in language while grounding key objects in the image regions they refer to. To train this behavior, we construct a scalable synthesis pipeline that distills correct visual reasoning traces, extracts the visual objects required by the traces, grounds them with a SAM3-based agent, and derives aligned point and box supervision from the resulting masks. We further propose grounding-aware reinforcement learning, which combines answer correctness rewards with dense grounding rewards that score whether generated object references match the correct image evidence. Across two counting benchmarks and four spatial reasoning benchmarks, adding visually grounded thinking to Gemma3-4B-IT consistently improves performance over the original model and the non-grounded thinking baseline. On spatial reasoning, the visually grounded thinking 4B models match, and in some cases surpass, Gemma3-27B-IT from the same model family. Our analysis shows that point grounding is well suited to counting, while box grounding benefits most from explicit grounding rewards on spatial tasks. Overall, our results show that VLMs think better when their intermediate thoughts are tied to the image regions that make them true.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.21959.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Minbyul Jeong&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; A working citation looks like proof -- but the fact that a link resolves does not mean the cited paper supports the claim. I find that current agentic models rarely fabricate citations (over 99% resolve), yet roughly 15.9% link to the wrong paper. Existing benchmarks miss this failure mode: when a question has a fixed answer key, a model can reproduce the expected source from that key rather than independently verifying that the source supports the claim. I introduce \openbiorq{}, a retrieval-grounded agentic benchmark of 12{,}553 unsolved biomedical research questions across 12 domains that treats open questions as a faithfulness-and-abstention probe. To my knowledge, this is the first biomedical benchmark to combine an agentic setting -- where the model must issue multiple tool calls -- with unsolved questions that have no answer key. Openness is verified against real follow-up evidence rather than a model's parametric knowledge. Difficulty is empirical: I anchor it on questions that three open-weight reference models fail to answer, rather than on subjective hardness labels. On this hardest subset, held-out models from the same lineage as the difficulty anchors solve only ~17%, while three independent frontier agents (Gemini-3-Pro, Opus-4.7, GPT-5.5) span a wide 29-60% range. The benchmark is thus hard, non-saturating (the best agent still leaves ~33-40\% unsolved), and discriminating across capability tiers. Beyond difficulty, I observe agentic collapse on the hardest questions, where agents stop using their tools. For the most collapse-prone model, blocking tool access entirely barely changes its score -- so tools stop paying off exactly where they are needed most. A frozen per-question checklist raises inter-judge agreement from Spearman 0.35 to 0.82.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.16122</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.21959</guid>
-      <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Context-Aware RL for Agentic and Multimodal LLMs</title>
+      <title>ABACUS: Adapting Unified Foundation Model for Bridging Image Count Understanding and Generation</title>
-      <link>https://arxiv.org/abs/2606.17053</link>
+      <link>https://arxiv.org/abs/2606.23835</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.17053.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Peiyang Xu, Bangzheng Li, Sijia Liu, Karthik R. Narasimhan, Pramod Viswanath, Prateek Mittal, Xingyu Fu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 10&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Large language models (LLMs) often fail when answering requires identifying a small but decisive piece of evidence within a long or complex context, such as a single line in a tool trace or a subtle detail in an image. We propose ContextRL, a context-aware reinforcement learning (RL) method that improves long-horizon reasoning and multimodal performance through an indirect auxiliary objective. Instead of supervising only the final answer, ContextRL presents the model with a query, an answer, and two highly similar contexts, and rewards it for selecting the context that supports the query--answer pair, thereby encouraging fine-grained grounding. We construct contrastive context data in two domains: for coding agents, trajectories serve as contexts, yielding 1k pairs built via condition filtering; for multimodal reasoning, images serve as contexts, yielding 7K pairs built via generative editing and similarity search. ContextRL achieves average gains of +2.2% over standard GRPO on 5 long-horizon benchmarks, and +1.8% across 12 diverse visual question answering benchmarks. To disentangle the effect of the proposed objective from that of additional data, we compare against data-augmentation baselines that repurpose the same contrastive contexts as standard query--context--answer examples. These baselines provide little to no improvement, showing that the gains arise from the proposed context-selection objective rather than from the contrastive data alone.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.23835.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Anindya Mondal, Sauradip Nag, Anjan Dutta&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 2&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; ABACUS is a unified vision-language model that handles object counting, crowd counting, referring-expression counting, and count-faithful image generation without any benchmark-specific training required. Our model is built on existing 3B-parameter unified foundation model and is adapted for object localization tasks using three key innovations: density-aware adaptive zooming with objectness maps for spatial grounding; a boundary-aware count policy via GRPO to eliminate crop-boundary errors; and a cycle-consistent GRPO strategy where the understanding branch self-critiques generated outputs, closing the understanding-generation gap without any external annotations. ABACUS achieves state-of-the-art results across seven benchmarks, outperforming both task-specific specialists and larger generalist models.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.17053</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.23835</guid>
-      <pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Mon, 22 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>LegalHalluLens: Typed Hallucination Auditing and Calibrated Multi-Agent Debate for Trustworthy Legal AI</title>
+      <title>GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents</title>
-      <link>https://arxiv.org/abs/2606.18021</link>
+      <link>https://arxiv.org/abs/2606.24551</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.18021.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Lalit Yadav, Akshaj Gurugubelli&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; AI systems deployed in legal workflows hallucinate at rates that aggregate metrics report at ~52%, but this average conceals where errors concentrate and in which direction they run, leaving compliance officers without an actionable signal for trustworthy deployment. We present LegalHalluLens, an auditing framework with three components: typed hallucination profiles across four legally-motivated claim categories (numeric, temporal, obligation/entitlement, factual) over CUAD (Hendrycks et al., 2021); a Risk Direction Index (RDI) that reduces omission-versus-invention bias to a single deployment-comparable scalar; and a typed debate pipeline calibrated to both magnitudes and directions. Across 510 contracts and 249,252 clause-level instances we measure a within-model gap of approximately 38-40 pp between obligation/numeric and temporal claims that aggregate reporting hides, and show that two systems with matched 52% rates can carry opposite RDIs. The debate pipeline reduces fabricated detections by 45% with per-category gains tracking the diagnosis, matching commercial APIs with a substantially smaller backbone (4B active parameters). Typed profiles and RDI surface failure modes that aggregate metrics hide; we further show these diagnostics serve as calibration inputs for multi-agent debate pipelines, where Skeptic challenges and asymmetric gates targeted at measured failure modes outperform generically-tuned debate. The framework supports direction-aware procurement, accountability, and agent design for legal AI deployed in the wild.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.24551.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Xiao Zhou, Siyue Zhang, Yilun Zhao, Jinbiao Wei, Tingyu Song, Arman Cohan, Chen Zhao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 25&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Computer-use agents can execute software tasks through either graphical interfaces or programmatic command interfaces, but existing evaluations confound interaction modality with differences in tasks, initial states, verifiers, and permitted actions. We introduce a matched execution-layer benchmark of 440 desktop tasks across 18 applications and 12 workflow categories, where screen-only GUI agents and skill-mediated CLI agents receive identical goals, states, and final-state verifiers while being restricted to modality-native actions. In this controlled setting, the strongest GUI agent reaches a 59.1% full pass rate, outperforming the strongest original-skill CLI agent at 48.2%; however, verifier-guided skill augmentation raises CLI success to 69.3%, showing that much of the CLI deficit comes from incomplete skill coverage rather than model capability alone. These results suggest that GUI and CLI expose different execution bottlenecks: GUI agents are limited by reliable grounded interaction over long-horizon workflows, whereas CLI agents are limited by the coverage and scalability of their skill interfaces.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.18021</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.24551</guid>
-      <pubDate>Tue, 16 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Mon, 22 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Adaptive Volumetric Mechanical Property Fields Invariant to Resolution</title>
+      <title>COrigami: An AI Pipeline for Co-Designing Flat-Foldable Visually Recognisable Origami</title>
-      <link>https://arxiv.org/abs/2606.18231</link>
+      <link>https://arxiv.org/abs/2606.26299</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.18231.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Rishit Dagli, Donglai Xiang, Vismay Modi, Xuning Yang, Gavriel State, David I. W. Levin, Maria Shugrina&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Accurate mechanical properties (or materials) Young's modulus (E), Poisson's ratio (ν) and density (ρ) are essential for reliable physics simulation of digital worlds, but most 3D assets lack this information. We propose AdaVoMP, a method for predicting accurate dense spatially-varying (E, ν, ρ) for input 3D objects across representations, improving the resolution, accuracy, and memory efficiency over the state-of-the-art. The foundation of our technique is a sparse and adaptive voxel structure SAV that efficiently represents both the input 3D shape and the material field output. We replace the fixed-voxel model of the most accurate prior method, VoMP, with a novel sparse transformer encoder-decoder model that learns to generate a unique SAV autoregressively for every input shape to represent its materials, achieving a resolution 16^3times higher than prior art. Experiments show that AdaVoMP estimates more accurate volumetric properties, even with lesser test-time compute than all prior art. This allows us to convert high-resolution complex 3D objects into simulation-ready assets, resulting in realistic deformable simulations.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26299.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Tom Zahavy, Shaobo Hou, Thomas Tumiel, James Doran, Francesco Faccio, Xidong Feng, Alex Havrilla, Igor Khytryi, Chenglei Li, Lisa Schut, Vivek Veeriah, Arijan Abrashi, Michał Kosmulski, Robert J. Lang, Nick Robinson, Brandon Wong, Marcus Chiam, Gloria Fang, Satinder Singh&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; While generative AI has achieved remarkable success in solving problems with verifiable solutions, generating physical art that satisfies both strict geometric constraints and subjective visual aesthetics remains a challenge. This paper presents an approach to tackle these difficulties in the domain of computational origami, a mathematically rigid environment that grounds artistic design within the equations of flat foldability. We present COrigami, an end-to-end AI-driven pipeline that assists the design cycle by generating crease patterns from natural language. Our pipeline involves generating a semantic stick figure, computing a base packing, solving for a flat-foldable crease pattern, shaping the flat-folded crease pattern, and refining the generated model using reinforcement learning driven by an autonomous aesthetic evaluation loop. Our system acts as a highly effective collaborative assistant, generating structural starting points that human artists can further expand and shape. By integrating algorithmic optimisation with autonomous aesthetic critique, this work demonstrates how AI systems can satisfy multi-objective physical constraints to enable reliable, mathematically grounded co-creativity.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.18231</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26299</guid>
-      <pubDate>Tue, 16 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Wed, 24 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>LooseControlVideo: Directorial Video Control using Spatial Blocking</title>
+      <title>Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents</title>
-      <link>https://arxiv.org/abs/2606.19495</link>
+      <link>https://arxiv.org/abs/2606.26080</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19495.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Shariq Farooq Bhat, Niloy J. Mitra, Kalyan Sunkavalli&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Precise 3D spatial orchestration in text-to-video generation remains a significant challenge, particularly for multi-object scenes where semantic layout and temporal dynamics are often entangled. While existing depth-conditioned models achieve good structural fidelity, they necessitate dense, frame-accurate guidance that is labor-intensive to author for dynamic events involving deformable objects. We present LooseControlVideo, a framework that enables intuitive and expressive control by using sparse, oriented 3D boxes as a "blocking" proxy. This allows users to author high-level layout and trajectory while leveraging a video generative model to generate realistic occlusions, dynamics and interactions. We achieve this by fine-tuning a Wan 2.2 backbone on a video dataset annotated with DNOCS, a novel encoding for 3D size, orientation and depth-ordered occlusions. Furthermore, our method allows for localized refinement, such as adjusting a jump trajectory or adding an interaction, with minimal disruption to the global scene context. Extensive evaluations on the nuScenes, HO-3D, and BEHAVE benchmarks demonstrate that LooseControlVideo significantly outperforms existing 2D-box and flow-based baselines. Our findings indicate a 1.2x to 3x improvement in Trajectory Error; 2x improvement in Rigid Motion Consistency; and a 1.5x to 2x increase in Occlusion Accuracy over current state-of-the-art layout-conditioned models, demonstrating that oriented 3D primitives provide good geometric prior for complex, multi-agent video authoring.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26080.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Changdae Oh, Wendi Li, Seongheon Park, Samuel Yeh, Tanwi Mallick, Sharon Li&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Process reward models enable fine-grained, step-level evaluation of LLMs, yet building them for agentic settings remains prohibitively difficult: long-horizon interactions, irreversible actions, and stochastic environment feedback make both human annotation and Monte Carlo estimation infeasible at scale. In this work, we show that reinforcement learning (RL) post-training already provides the ingredients for effective step-level scoring, eliminating the need for dedicated reward model training altogether. Concretely, we derive an implicit advantage under a general stochastic Markov decision process, which we term progress advantage -- log-probability ratio between the RL-trained policy and its reference policy exactly recovers the optimal advantage function. This formulation makes the resulting signal annotation-free, domain-agnostic, and available as a byproduct of the standard RL post-training pipeline. We validate the effectiveness of the progress advantage across three different applications: test-time scaling, uncertainty quantification, and failure attribution on five benchmarks and four model families. Across all settings, it consistently outperforms confidence-based baselines and, despite requiring no task-specific training, surpasses dedicated trained reward models. We complement these results with deeper analyses on characteristics of progress advantage, offering practical guidance for adoption in real-world agentic systems.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19495</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26080</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Wed, 24 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Configurable Clinical Information Extraction with Agentic RAG: What Works, What Breaks, and Why</title>
+      <title>Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It</title>
-      <link>https://arxiv.org/abs/2606.19602</link>
+      <link>https://arxiv.org/abs/2606.26027</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19602.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Osman Alperen Çinar-Koraş, Marie Bauer, Sameh Khattab, Merlin Engelke, Moon Kim, Stephan Settelmeier, Shigeyasu Sugawara, Fabian Freisleben, Felix Nensa, Jens Kleesiek&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Patient contexts span hundreds of heterogeneous documents and thousands of structured data points, yet the document-level metadata that AI systems need for retrieval and triage is absent or incomplete. Standard retrieval-augmented generation fails on this data, mishandling temporal reasoning, cross-document dependencies, and missing metadata. We deploy ACIE (Agentic Clinical Information Extraction) at University Medicine Essen: an on-premise agentic RAG pipeline that reasons over complete patient contexts and grounds every answer in source passages for clinician verification. We quantify the metadata gap, trace the architectural decisions it shaped, and evaluate extraction alongside an independent retrospective lymphoma registry study, in which nuclear-medicine physicians verify every extracted value against its cited sources. Across 7,326 judgments, clinicians accepted 96.5\% of extractions, with per-type acceptance ranging from 80\% to 99\%.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26027.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yupu Hao, Zhuoran Jin, Huanxuan Liao, Kang Liu, Jun Zhao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 14&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Tool use enables large language models (LLMs) to perform complex tasks, and recent agentic reinforcement learning (RL) methods show promise for enhancing model capabilities. However, RL alone often leads to instability or limited gains in tool-use tasks. In our experiments, some models exhibit catastrophic collapse, where performance abruptly drops and tool-invocation structures fail. The analysis reveals that these failures stem from unexpected probability spikes in specific control tokens, disrupting structured execution, yet the underlying tool-use capability remains intact, merely obscured by specific formats. To address this, we systematically investigate a diverse set of supervisory signals, including off-policy supervision, hint-based guidance, erroneous example supervision, and others, applied under both synchronous and interleaved training schemes. We find that interleaving supervised fine-tuning (SFT) with RL substantially improves stability, but exhibits degraded performance under format and content out-of-distribution (OOD) evaluation. We also analyze the impact of learning rates and generalization across settings. These results highlight the importance of understanding RL failures and demonstrate how diverse supervisory signals can guide exploratory learning, enabling robust training of LLMs for complex, multi-step tool-use tasks. Our Code is available at https://github.com/hypasd-art/Tool-RL-Box.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19602</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26027</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Wed, 24 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States</title>
+      <title>Fast LeWorldModel</title>
-      <link>https://arxiv.org/abs/2606.19334</link>
+      <link>https://arxiv.org/abs/2606.26217</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19334.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Denis Peskoff, Joe Barrow, Christopher Vu, Diag Davenport&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Progress in legal AI increasingly depends on access to authoritative legal text at scale. Yet one of the most consequential layers of American law remains largely absent from existing machine-readable corpora: local ordinances. Local codes govern zoning, housing, business licensing, public health, noise, animal control, and many other domains of everyday regulation, but they are fragmented across vendor platforms designed for human browsing rather than bulk research access. We introduce LOCUS - the Local Ordinance Corpus for the United States - a comprehensive corpus and county-harmonized access layer for U.S. municipal and county ordinance codes. The raw corpus, available for release to researchers, represents nearly all publicly available municipal and county ordinance codes. The resulting raw corpus contains codes from 9,239 cities and counties. A smaller county-harmonized LOCUS access layer provides coverage for the largest 2,309 of 3,144 U.S. counties, accounting for a majority of the population. We use OCR to handle the myriad of document formats that have kept the law from being a public resource. We release the corpus with coverage metadata to support reproducibility, downstream legal AI research, and the incremental expansion of machine-readable access to local law. We train a collection of ModernBERT-based classifiers and scorers to facilitate analyzing U.S. local law among several dimensions, such as opacity and paternalism, that have not previously been studied at this scale. LOCUS-v1 and its derivative models are available at: https://huggingface.co/datasets/LocalLaws/LOCUS-v1&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26217.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yuntian Gao, Xiangyu Xu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 20&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Joint-Embedding Predictive Architectures (JEPAs), including recent LeWorldModel (LeWM), have become a promising foundation for reconstruction-free visual world models. For visual planning, however, LeWM evaluates candidate action sequences by repeatedly applying a local one-step latent transition model. This autoregressive rollout makes planning computationally expensive and exposes the predicted trajectory to accumulated latent errors as the horizon grows. We propose Fast LeWorldModel (Fast-LeWM), a fast latent world model that replaces repeated local rollout with action-prefix prediction. Given the current latent and a candidate action sequence, Fast-LeWM encodes its prefixes and predicts the future latents reached after executing those prefixes in parallel. By making action prefixes the basic prediction unit, Fast-LeWM directly models action effects accumulated to different extents over multiple horizons. This prefix-level supervision forces the model to learn how states continuously evolve under different action prefixes, rather than only fitting one-step state transitions. During planning, the predictor can use the last prefix token from the encoded action sequence to evaluate the corresponding future latent without explicitly rolling through each intermediate imagined state. Across multiple tasks, Fast-LeWM improves average success over LeWM while substantially reducing planning time, achieving lower open-loop latent loss whose growth becomes significantly slower as the rollout horizon increases.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19334</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26217</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Wed, 24 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>FAPO: Fully Autonomous Prompt Optimization of Multi-Step LLM Pipelines</title>
+      <title>The Verification Horizon: No Silver Bullet for Coding Agent Rewards</title>
-      <link>https://arxiv.org/abs/2606.19605</link>
+      <link>https://arxiv.org/abs/2606.26300</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19605.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Paul Kassianik, Baturay Saglam, Huaibo Zhao, Blaine Nelson, Supriti Vijay, Aman Priyanshu, Amin Karbasi&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 10&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Multi-step LLM pipelines fail through interactions among retrieval, reasoning, and formatting steps, so prompt-only optimization can miss bottlenecks in the chain. We present FAPO (Fully Autonomous Prompt Optimization), a framework that lets Claude Code optimize an LLM pipeline inside a standardized codebase. FAPO evaluates a pipeline, inspects intermediate steps, diagnoses failures, proposes scoped changes, and validates variants repeatedly to optimize against a score function. It first tries prompt edits and, only when prompt optimization appears insufficient, changes chain structure within the permitted scope when attribution identifies a structural bottleneck. Across six benchmarks and three task models, FAPO beats the baseline GEPA in 15 of 18 model-benchmark comparisons. In 11 model-benchmark comparisons, FAPO wins with non-overlapping mean pm trial-standard-deviation ranges, and the mean FAPO-GEPA gain is +14.1 pp. In the six HoVer and IFBench comparisons where prompt-first search escalated to structural changes, FAPO wins all six with a mean gain of +33.8 pp. FAPO also improves performance on security tasks: on CTIBench-RCM, a security CVE-to-CWE task, prompt-only FAPO lifts test accuracy by +4.0 pp on GPT-5, +7.1 pp on Foundation-Sec-8B-Instruct, and +2.0 pp on Foundation-Sec-8B-Reasoning. These results position FAPO as a state-of-the-art pipeline optimization technique for both general-purpose and security-focused tasks.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26300.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Binghai Wang, Chenlong Zhang, Dayiheng Liu, Jiajun Zhang, Jiawei Chen, Mouxiang Chen, Rongyao Fang, Siyuan Zhang, Xuwu Wang, Yuheng Jing, Zeyao Ma, Zeyu Cui&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 36&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; A classical intuition holds that verifying a solution is easier than producing one. For today's coding agents, this intuition is being inverted: as foundation models develop stronger reasoning capabilities and engineering harnesses grow more sophisticated, generating complex candidate solutions is no longer difficult -- reliably verifying them has become the harder problem. Every verifier we can build is only a proxy for human intent, never the intent itself. This makes verification subject to a twofold difficulty: first, intent is underspecified by nature, making it inherently hard to faithfully check whether it has been fulfilled; second, during model training, optimization widens the gap between proxy and intent -- manifesting as reward hacking or signal saturation. To address this, we characterize the quality of verification signals along three dimensions -- scalability, faithfulness, and robustness -- and argue that achieving all three simultaneously is the central challenge. We further study four reward constructions: a test verifier for general coding tasks, a rubric verifier for frontend tasks, the user as verifier for real-world agent tasks, and an automated agent verifier for long-horizon tasks. Across different task types and policy capability levels, we conduct in-depth analysis and experiments on the core challenges of reward design and how to more effectively leverage reward signals. Experiments show that targeted verification design can effectively suppress reward hacking, improve task completion quality, and achieve significant gains across multiple internal and public benchmarks. These experiences collectively point to a core observation: no fixed reward function can remain effective as policy capability continues to grow; and verification must co-evolve with the generator.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19605</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26300</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Wed, 24 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>ImageWAM: Do World Action Models Really Need Video Generation, or Just Image Editing?</title>
+      <title>EO-WM: A Physically Informed World Model for Probabilistic Earth Observation Forecasting</title>
-      <link>https://arxiv.org/abs/2606.19531</link>
+      <link>https://arxiv.org/abs/2606.27277</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19531.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yuyang Zhang, Wenyao Zhang, Zekun Qi, He Zhang, Haitao Lin, Jingbo Zhang, Yao Mu, Xiaokang Yang, Wenjun Zeng, Xin Jin&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 14&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; World Action Models (WAMs) commonly rely on video generation to bridge visual world modeling and robot control. However, video-based WAMs face three coupled limitations: dense multi-frame future tokens make inference costly, full video prediction spends capacity on action-irrelevant temporal and appearance details, and long-horizon future imagination may introduce errors that mislead action prediction. These issues raise a simple question: Does world action model really need video generation? We propose ImageWAM, a simple WAM framework that repurposes pretrained image editing models for robot action prediction. In contrast to video generation, image editing provides a better-matched prior: it only needs to model a target-frame transformation, focuses on action-relevant current-to-target visual differences, and grounds task instructions to localized visual changes through edit pretraining. In practice, ImageWAM does not decode the target frame at inference time; instead, it conditions a flow-matching action expert on the KV caches produced by image-editing denoising, using them as a compact world-action context. ImageWAM outperforms standard VLA baselines and matching competitive WAMs without additional policy pretraining across different simulator and real-world experiments. It also reduces FLOPs to 1/6 and latency to 1/4 of video-based WAMs. Attention analysis further shows that editing caches focus on task-relevant change regions, supporting image editing as an effective alternative to video-based world-action modeling.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27277.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Junwei Luo, Shuai Yuan, Zhenya Yang, Yansheng Li, Zhe Liu, Hengshuang Zhao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 2&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Earth Observation (EO) forecasting aims to predict future Earth surface dynamics from satellite observations under changing meteorological conditions. In this paper, we view this task as a partially observed, weather-driven world modeling problem, in which weather acts as a conditioning signal, while forecasting remains uncertain due to sparse observations and unobserved land-surface states. However, existing methods do not fully capture this setting: deterministic models collapse uncertainty into a single future prediction, while diffusion-based methods typically treat weather variables as undifferentiated conditioning signals, and existing benchmarks focus mainly on reconstruction accuracy rather than whether forecasts respond correctly to changed weather forcing.We introduce EO-WM, a video diffusion transformer for multispectral EO forecasting. EO-WM incorporates a physically informed conditioning framework that represents meteorological forcing through a climatological baseline, weather anomalies, and cumulative physical stress signals. Specifically, it separates baseline and anomaly through distinct conditioning pathways, and accumulates anomalous forcing over time to capture sustained heat and drought stress. To evaluate weather-response behavior beyond standard metrics, we introduce two diagnostic benchmarks: an Extreme Summer Benchmark for severity-aware prediction of vegetation degradation under extreme weather, and a Seasonal Matched-Pair Benchmark for testing response fidelity under changed weather forcing. Experiments show that EO-WM reduces the error in predicted Normalized Difference Vegetation Index (NDVI) decline amplitude by a relative 5.63% and improves directional hit rate by a relative 7.80%, while remaining competitive on standard pixel-level metrics. The benchmarks and model will be made open-source at https://github.com/Luo-Z13/EO-WM.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19531</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27277</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Playful Agentic Robot Learning</title>
+      <title>When Does Combining Language Models Help? A Co-Failure Ceiling on Routing, Voting, and Mixture-of-Agents Across 67 Frontier Models</title>
-      <link>https://arxiv.org/abs/2606.19419</link>
+      <link>https://arxiv.org/abs/2606.27288</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19419.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Junyi Zhang, Jiaxin Ge, Hanjun Yoo, Letian Fu, Zihan Yang, Yaowei Liu, Raj Saravanan, Shaofeng Yin, Justin Yu, Dantong Niu, Zirui Wang, Roei Herzig, Ken Goldberg, Yutong Bai, David M. Chan, Ion Stoica, Angjoo Kanazawa, Jiahui Lei, Haiwen Feng, Trevor Darrell&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 43&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Current agentic robot systems can write executable Code-as-Policy programs, observe feedback, and revise behavior across multiple attempts, but they remain largely task-driven: reusable skills are acquired only after explicit instructions. We study Playful Agentic Robot Learning, where an embodied coding agent uses self-directed play as a continual skill-learning stage before downstream tasks arrive. We introduce RATs, Robotics Agent Teams designed for play-time skill acquisition. During play, RATs proposes novel yet learnable exploratory tasks, plans and executes robot-code policies, verifies intermediate progress, diagnoses failures, retries with dense, step-level feedback, and distills successful executions into a persistent code skill library. At test time, the agent reuses relevant skills from this frozen library to help solve new tasks. Experiments in LIBERO-PRO and MolmoSpaces show that play-learned skills improve held-out downstream tasks over no-play and random-play baselines, with 20.6 and 17.0 percentage-point gains over CaP-Agent0 on LIBERO-PRO and MolmoSpaces, respectively. Moreover, the learned skills can be plugged into other inference-time Code-as-Policy agents by simply retrieving them into the context, improving RoboSuite and real-world transfer by 8.9 and 8.8 points, respectively, without finetuning the underlying model.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27288.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Josef Chen&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Multi-model LLM systems such as routing, voting, cascades, fusion, and mixture-of-agents are used to beat single-model accuracy. We show that their gain is capped by a quantity the field rarely reports. For any policy whose output is one member model answer, accuracy cannot exceed one minus beta, where beta is the rate at which every model is wrong on the same query. In contrast, the usual diagnostic, average pairwise error correlation rho, cannot identify beta: error laws with identical marginals and pairwise correlations can have different all-wrong rates. A Clopper-Pearson bound on beta gives a finite-sample certificate on the largest gain any router, vote, or cascade could deliver before training a router.   Across 67 models from 21 providers, a tetrachoric-calibrated single-factor model still underprices the all-wrong tail: on open-ended mathematics, observed beta is 0.052 versus 0.023 under the full 67-model Gaussian copula, about 2.5 times underpricing, with 90 percent CI 1.7 to 3.4 and k equals 17. The effect recurs on execution-graded code, where beta is 0.079. Re-asking the same GPQA-Diamond questions in free-response rather than multiple-choice form reopens the tail, with beta 0.127 and a five-judge panel with kappa 0.73 to 0.92, locating co-failure in answer format rather than subject. At matched quality, low-rho heterogeneous ensembles beat high-rho Self-MoA, but on checkable tasks in our pool, combining models rarely beats the single best model without a strong query-level routing signal. Gains come from models failing on different questions, not from adding more models.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19419</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27288</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance</title>
+      <title>Information-Aware KV Cache Compression for Long Reasoning</title>
-      <link>https://arxiv.org/abs/2606.19195</link>
+      <link>https://arxiv.org/abs/2606.26875</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19195.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Kangsheng Duan, Ziyang Xu, Wenyu Liu, Xiaohu Ruan, Xiaoxin Chen, Xinggang Wang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 113&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; While 10B-level industrial foundation models have pushed the boundaries of image inpainting, their prohibitive computational costs severely hinder practical deployment. Constructing a highly optimized task-specific specialist offers a promising solution; however, extreme structural compression inevitably triggers a severe representation bottleneck. To conquer this, we propose Moebius, a highly efficient lightweight inpainting framework. We systematically reconstruct the diffusion backbone by introducing the Local-λ Mix Interaction (LλMI) block. Comprising Local-λ and Interactive-λ modules, it elegantly summarizes spatial contexts and global semantic priors into fixed-size linear matrices, preserving complex latent interactions while drastically shedding parameters. Furthermore, to unlock the full representational capacity of this highly compact architecture, we synergistically pair it with an adaptive multi-granularity distillation strategy. Operating strictly within the latent space to avoid expensive pixel-space decoding, this strategy dynamically balances multiple gradient-based losses to achieve high-fidelity alignment. Extensive experiments across natural and portrait benchmarks demonstrate that this optimal synergy enables Moebius to rival or even surpass the generation quality of the 10B-level industrial generalist FLUX.1-Fill-Dev. Remarkably, Moebius achieves this using less than 2\% of the parameters (0.22B vs. 11.9B) while delivering a &gt;15times acceleration in total inference time, setting a new efficiency standard for high-fidelity inpainting. Project page at https://hustvl.github.io/Moebius.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26875.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Jushi Kai, Zhuiri Xiao, Alexandra Birch, Zhouhan Lin&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Reasoning capability has advanced rapidly in large language models (LLMs), leading to an increasing size of key-value (KV) cache in both prefilling and decoding stages. Existing KV cache compression methods mainly rely on attention weights to estimate token importance. While attention effectively captures contextual relevance, it overlooks complementary information-theoretic signals related to predictive uncertainty and token informativeness. In this paper, we revisit token importance from a forward-looking perspective and introduce Forward Influence, a metric that measures how compressed tokens affect future contexts. Our analysis reveals that tokens selected by attention scores mainly influence nearby contexts, whereas tokens associated with high predictive uncertainty exhibit substantially stronger influence on distant future contexts. Based on the observation, we propose InfoKV, an entropy-aware KV cache compression framework that incorporates information-theoretic signals. It combines token-level predictive uncertainty with layer-wise representation evolution and integrates the resulting entropy scores with attention scores during reasoning. Experiments on long-context reasoning benchmarks with Llama-3.1, Llama-3.2, and DeepSeek-R1 demonstrate that InfoKV consistently outperforms existing attention-based KV compression methods in both long prefilling and decoding scenarios.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19195</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26875</guid>
-      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>JAMER: Project-Level Code Framework Dataset and Benchmark on Professional Game Engines</title>
+      <title>Hallucination in World Models is Predictable and Preventable</title>
-      <link>https://arxiv.org/abs/2606.19830</link>
+      <link>https://arxiv.org/abs/2606.27326</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19830.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Jianwen Sun, Chuanhao Li, Zizhen Li, Yukang Feng, Fanrui Zhang, Yifei Huang, Yu Dai, Kaipeng Zhang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 3&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Current AI-driven game development has made substantial progress in asset generation, gameplay design, and web-based game coding, yet project-level code engineering on professional game engines remains largely unexplored due to the absence of large-scale datasets and deterministic evaluation methods. We present JamSet and JamBench, the first project-level game code framework dataset and benchmark built on a professional game engine. Our key insight is that Game Jam competitions, community events where developers build complete games under tight time constraints, yield thousands of open-source projects suitable for this purpose. Building on the Godot engine's text-based format and headless execution mode, we design a deterministic verification pipeline from file integrity to runtime behavior collection, distilling 8,133 verified projects from over 240,000 repositories. Of these, 300 manually verified projects form JamBench; the rest constitute JamSet. JamBench defines theme-driven generation and code completion tasks, evaluated through a pipeline combining compilation pass rates, Structural Completeness Score (SCS), and Behavioral Alignment Score (BAS). Evaluation of 9 frontier models reveals a capability cliff as project scale increases, with runtime pass rates dropping from 80.4% on small projects to 5.7% on large ones (Task2a). Code Agents improve compilation rates yet yield no gains in runtime behavioral quality, indicating that the bottleneck lies in architectural design rather than syntactic correctness. Experiments validate JamSet as effective training data. All data and code are publicly available.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27326.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Nicklas Hansen, Xiaolong Wang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 7&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Modern generative world models render increasingly realistic action-controllable futures, yet they frequently hallucinate: rollouts remain visually fluent while drifting from the ground-truth dynamics. We hypothesize that hallucination concentrates in low-coverage regions of the state-action space, where lightweight data-centric signals can both detect it and guide mitigation. To test this, we introduce MMBench2, a 427-hour, 210-task dataset for visual world modeling with ground-truth actions, rewards, and live simulators, and train a 350M-parameter world model on it. We identify three distinct hallucination modes: perceptual, action-marginalized, and scene-diverging -- each anchored to a different stage of the pipeline, and develop three signals that accurately predict where the model will fail. To close coverage gaps at training time, we develop a coverage-aware sampling technique; to close them online, our hallucination predictors serve as curiosity rewards for targeted data collection, yielding a data-efficient finetuning recipe that adapts the pretrained world model to entirely unseen environments with as few as 50 real environment trajectories. Overall, our findings reveal that hallucination in world models is inherently a data coverage issue, and that the same signals used to detect it can also be used for mitigation.   An interactive web version of our paper is available at https://www.nicklashansen.com/mmbench2&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19830</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27326</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>The FID Lottery: Quantifying Hidden Randomness in Generative-Model Evaluation</title>
+      <title>PhysiFormer: Learning to Simulate Mechanics in World Space</title>
-      <link>https://arxiv.org/abs/2606.20536</link>
+      <link>https://arxiv.org/abs/2606.27364</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20536.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Nicolas Dufour, Alexei A. Efros, Patrick Pérez&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 4&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; The Frechet Inception Distance (FID) is the de facto arbiter of image generation, yet most papers report just a single number from a single trained model using a single sampling seed. How reproducible is that number if we retrain the model, or merely resample from it? In this paper, we treat FID as a random variable on a two-axis panel of training and generation seeds, and measure its variance directly on several hundred SiT networks trained on class-conditional ImageNet 256x256. We report surprising findings: (a) Retraining the model using the same recipe with a different seed moves FID 3.2x more (in Inception feature space) than redrawing samples from a fixed network. (b) That gap is driven by three factors: random initialisation, data ordering, and the per-step Gaussian noise of the flow-matching loss. (c) Increasing compute or model size barely tightens the spread, holding the FID coefficient of variation (CoV) inside a 1-2% band. (d) Per-cell classifier-free-guidance tuning halves the spread but reshuffles which seeds work best, and a lucky training seed reaches the same FID with up to 2x less compute than an unlucky one. Based on these findings, we recommend a new FID evaluation protocol: evaluate under per-cell optimal guidance, treat any FID gap below the empirically measured ~1.3% CoV as inconclusive, and report an error bar over several training seeds rather than a single FID number.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27364.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yiming Chen, Yushi Lan, Andrea Vedaldi&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 9&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; We present PhysiFormer, a diffusion transformer for physically-plausible 3D object motion. Unlike video world models that operate in view-dependent pixel space, PhysiFormer represents objects as 3D meshes expressed in world coordinates. Given the initial vertex positions and velocities, as well as object material type, rigid or elastic, the model samples future vertex trajectories. While related neural physics approaches build on ad-hoc latent spaces or explicitly enforce rigidity and causality, PhysiFormer shows that excellent results can be obtained without any such inductive biases, by casting vertex trajectory prediction as a single denoising diffusion process directly in world coordinates. The probabilistic formulation captures uncertainty in the learned dynamics, enabling diverse plausible futures from initial conditions, making this framework potentially useful for applications with unobserved uncertainty. The model features attention factorised over time, space, and objects for efficiency, enabling permutation-invariant multi-object reasoning without needing explicit object encoding. Trained on over 100k simulated trajectories, PhysiFormer generates rigid and elastic mechanics, and generalises to mixed-material settings, unseen real-world geometries, and larger object counts. It substantially outperforms autoregressive baselines in trajectory accuracy, rigidity preservation, and momentum-based physical consistency. Our results position coordinate-space diffusion as a promising step toward view-invariant, geometry-aware world modelling for robotics, graphics, and physical design. Visualisations, code, and models are available at https://yimingc9.github.io/physiformer.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20536</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27364</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Holo-World: Unified Camera, Object and Weather Control for Video World Model</title>
+      <title>Confidence-Aware Tool Orchestration for Robust Video Understanding</title>
-      <link>https://arxiv.org/abs/2606.20083</link>
+      <link>https://arxiv.org/abs/2606.26904</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20083.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Xiangchen Yin, Wenzhang Sun, Jiahui Yuan, Zijie Liu, Yinda Chen, Wei Li, Dachun Kai, Chunfeng Wang, Xiaoyan Sun&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 5&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Video world models are moving toward preserving an observed world under controllable camera and object motion while allowing its environmental state to change. Yet these controls remain isolated, and weather generation typically relies on a source video or reconstructed scene that already specifies future structure. We study a first-frame-anchored source-to-state setting, where the model starts from a single image and follows explicit camera and object controls and an optional weather instruction, then generates a video that either preserves the source world or transfers it to a target weather state. To address these challenges, we first build HoloStateData, a state video dataset that turns diverse videos into unified control samples for camera, object, and weather supervision. Second, we introduce Holo-World, a unified controllable video world model that jointly controls scene from a single image. Its Unified Scene Adapter factorizes world preservation and weather transfer into distinct parameter subspaces, using rendered background, geometry buffers, and object controls to maintain controlled scene structure while modeling weather-dependent appearance and particle effects. Additionally, Scene-Weather Decomposed CFG guides scene and weather residuals separately, strengthening target weather effects without over-amplifying the full condition. Quantitative and qualitative experiments demonstrate that Holo-World maintains precise camera and object control with consistent scene structure while transferring scenes into diverse target weather state, outperforming video-to-video weather editing baselines on weather-state generation. Our project page is available at https://xiangchenyin.github.io/Holo-World/.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26904.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yangfan He, Yujin Choi, Jaehong Yoon&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 9&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Video reasoning language models implicitly assume that every input frame is equally reliable. This leads to what we term the Blind Trust Problem: under realistic perturbations such as motion blur, glare, or occlusion, frontier video reasoning models can suffer 15-30%p accuracy drops on real-world embodied benchmarks, while remaining unaware that their visual evidence has been degraded. To address this challenge, we propose Robust-TO, an agentic video understanding framework that explicitly integrates per-frame trustworthiness into every stage of reasoning. Robust-TO organizes heterogeneous visual perception tools under a unified evidence interface. Each tool receives a sub-query derived from the original question and a set of trustworthy frames selected by the reliability-relevance score. It returns evidence in a shared format: a concrete prediction (e.g., a bounding box, motion trajectory, recognized text, or action label), temporal grounding, and a calibrated reliability score. During reasoning, these calibrated scores guide evidence weighting in a three-tier synthesis process (high/medium/low) and define a confidence-cost GRPO reward that jointly optimizes correctness, evidence reliability, and efficiency. On two video reasoning benchmarks spanning eight tasks, Robust-TO achieves 56.4% average accuracy on clean inputs, surpassing the strongest open-source baseline by 10.6%p and outperforming Gemini-2.5-Pro (46.2%). Under five realistic corruption types, Robust-TO maintains 54.3% average accuracy, 5.8%p above the strongest open-source baseline, while exhibiting the smallest clean-to-corrupted accuracy drop among all compared methods.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20083</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26904</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>HumanScale: Egocentric Human Video Can Outperform Real-Robot Data for Embodied Pretraining</title>
+      <title>LISA: Likelihood Score Alignment for Visual-condition Controllable Generation</title>
-      <link>https://arxiv.org/abs/2606.20521</link>
+      <link>https://arxiv.org/abs/2606.27192</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20521.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Juncheng Ma, Jianxin Bi, Yufan Deng, Xuanran Zhai, Kewei Zhang, Ye Huang, Bo Liang, Shukai Gong, Jiankai Tu, Xiaotian Tang, Jiaxin Li, Kaiqi Chen, Duomin Wang, Yuqi Wang, Bingyi Kang, Eric Huang, Zhiyang Dou, Zhen Dong, Enze Xie, Wojciech Matusik, Tat-Seng Chua, Daquan Zhou&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Embodied foundation models are expected to benefit from data scaling like large language models, but face a much tighter data bottleneck. Teleoperated real-robot trajectories remain the dominant pretraining source due to their precise action supervision and embodiment alignment, yet their scalability is limited by high collection cost, acquisition difficulty, and low behavioral and environmental diversity. These limitations have sparked interest in egocentric human video as a scalable, substantially lower-cost, and more diverse alternative for embodied model pretraining. However, its effectiveness compared to teleoperated real-robot data remains underexplored. To address this question, we conduct a systematic study comparing egocentric human video and teleoperated real-robot trajectories as pretraining data sources for embodied foundation models, under fixed post-training and validation protocols. Surprisingly, we find that egocentric data, when processed through a carefully designed filtering and labeling pipeline, is not merely a viable substitute for model pretraining but can lead to superior performance. With the same amount of pretraining data, models pretrained on egocentric data achieve a 24% lower validation loss on real-robot action prediction, as well as 52.5% and 90% higher success rates on in-distribution and out-of-distribution real-robot task execution, respectively. This finding verifies a scalable paradigm for embodied foundation models: pretrain on egocentric human video to learn diverse world representations, then adapt with a small amount of labeled real-robot data for action-space alignment. We hope this study encourages broader exploration of egocentric data and offers guidance for data quality assessment before costly robot data collection.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27192.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yanghao Wang, Hongxu Chen, Jiazhen Liu, Zhenqi He, Rui Liu, Zhen Wang, Long Chen&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 12&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; The prevalent dual-branch paradigm, i.e., training a side network to encode visual conditions and fusing its intermediate-layer features to a frozen pretrained main network, has shown remarkable success in visual-condition controllable generation. Despite its widespread adoption, the role of the side branch and its training efficiency remain underexplored. In this paper, we first revisit this mainstream paradigm through the lens of score-based generative modeling: 1) The main network preserves visual perceptual quality by providing a prior unconditional score. 2) The side network steers conditional control by implicitly contributing a likelihood score. Guided by this perspective, we propose LIkelihood Score Alignment (LISA), an effective regularization method that explicitly aligns the intermediate feature of the side network with an approximated likelihood score. Specifically, we first hook features from a designated layer of the side network and project them into the score latent space by a lightweight decoder. Then, we construct an approximated likelihood score target and calculate the distance between the decoder's output and this target as an additional regularization loss. Finally, we jointly optimize the side network and decoder with both standard diffusion loss and our regularization loss. Experiments across various image/video tasks, architectures, and diffusion/flow models demonstrated that LISA can not only consistently accelerate the training convergence and improve final synthetic results, but also encourage the side network's features to be more disentangled for conditional modeling with negligible additional training cost and zero extra inference cost.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20521</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27192</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Rethinking Shrinkage Bias in LLM FP4 Pretraining: Geometric Origin, Systemic Impact, and UFP4 Recipe</title>
+      <title>Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments</title>
-      <link>https://arxiv.org/abs/2606.20381</link>
+      <link>https://arxiv.org/abs/2606.14397</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20381.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Qian Zhao, Kunlong Chen, Changxin Tian, Zhonghui Jiang, Haitao Zhang, Chaofan Yu, Peijie Jiang, Mingliang Gong, Jia Liu, Ziqi Liu, Zhiqiang Zhang, Jun Zhou&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; FP4 training promises substantial reductions in memory and computation cost for LLM pretraining, yet current FP4 hardware paths and recipes, including NVIDIA Blackwell/Rubin-class systems and AMD MI350-series GPUs, remain centered on E2M1 data elements. In this study, we identify a fundamental limitation of that choice: non-uniform formats such as E2M1 inherently suffer from Shrinkage Bias, a systematic negative rounding error caused by the geometric asymmetry of their representable bins. We show that this bias accumulates multiplicatively across layers and is amplified by the Random Hadamard Transform (RHT), providing a unified explanation for the training instability observed in existing E2M1-based FP4 recipes. In contrast, uniform grids (E1M2/INT4) bypass this grid-geometry error and better convert the improved bucket utilization from RHT into higher quantization quality. Based on this finding, we propose UFP4, a uniform 4-bit training recipe that applies RHT to all three training GEMMs while restricting stochastic rounding to dY alone. On Dense 1.5B, MoE 7.9B, and MoE 124B long-run pretraining, UFP4 consistently achieves lower BF16-relative loss degradation than strong E2M1-based baselines, supported by scaling-law analysis and ablation studies. Our results suggest that future accelerators should support E1M2/INT4-style uniform 4-bit grids as first-class training primitives alongside E2M1.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.14397.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Mykola Vysotskyi, Runqi Lin, Grzegorz Biziel, Michal Zakrzewski, Sebastian Montagna, Damian Rynczak, Shreyansh Padarha, Kumail Alhamoud, Zihao Fu, William Lugoloobi, Kai Rawal, Hanna Yershova, Xander Davies, Taras Rumezhak, Guohao Li, Fazl Barez, Baoyuan Wu, Arkadiusz Drohomirecki, Yarin Gal, Chris Russell, Christopher Summerfield, Adam Mahdi, Volodymyr Karpiv, Philip Torr, Adel Bibi&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 14&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; As agentic systems continue to evolve and are widely deployed in real-world scenarios, there is a growing demand to faithfully evaluate their capabilities. However, current benchmarks are typically built on popular applications with relatively simple tasks and focus on a narrow set of capabilities while overlooking broader dimensions, resulting in saturated performance on modern agents and failing to probe their limitations. To this end, we introduce GauntletBench, a web-based benchmark for evaluating agent generalisation in challenging scenarios, focusing on three underexplored capabilities (temporal perception, graphical understanding, and 3D reasoning), across five less-covered professional applications (Video Editor, Workflow Builder, 3D Modeller, Flight Analyser, and Circuit Designer), each with 20 vision-intensive tasks (100 in total). Our benchmark provides a modular pipeline that comprises an environment compatible with both open- and closed-source agent frameworks, a controlled web-based application, a well-structured task suite, and an automated evaluation engine with diverse metrics. Contrary to widespread expectations, our empirical results reveal that frontier agentic systems remain far from achieving human-level performance. Even the state-of-the-art agent achieves only a 19.1% success rate on our GauntletBench, highlighting the limitations in these overlooked capabilities and generalisation. By comparison, non-expert human annotators achieve over 80% success on our challenging yet feasible tasks, revealing the substantial gap between current agent capabilities and those required for complex real-world scenarios.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20381</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.14397</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>LedgerAgent: Structured State for Policy-Adherent Tool-Calling Agents</title>
+      <title>JetSpec: Breaking the Scaling Ceiling of Speculative Decoding with Parallel Tree Drafting</title>
-      <link>https://arxiv.org/abs/2606.20529</link>
+      <link>https://arxiv.org/abs/2606.18394</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20529.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Md Nayem Uddin, Amir Saeidi, Eduardo Blanco, Chitta Baral&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 6&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Policy-adherent tool-calling agents in customer-service domains must maintain task states across turns while calling tools and obeying domain policies. Task states consist of relevant facts, identifiers, constraints, and conditions observed through user interaction and tool calls. In standard agents, task states are not represented separately. Observations, tool returns, and policy instructions are placed in the prompt, leaving agents to reconstruct the relevant states from the prompt each time they decide what to do next. This design makes state management implicit, creating two common failure modes. An agent may retrieve the right facts but later ground its decision in stale, missing, or incorrect information; and a syntactically valid tool call may still violate a domain policy that depends on the current task state. We introduce LedgerAgent, an inference-time method for tool-calling agents that maintains observed task states in a separate ledger and renders the states into the prompt. The ledger is also used to check state-dependent policy constraints before environment-changing tool calls are executed, blocking policy violations. Across four customer-service domains and a mixed panel of open- and closed-weight models, LedgerAgent improves average passk over a standard prompt-based tool-calling approach, with the largest gains under stricter multi-trial consistency metrics.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.18394.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Lanxiang Hu, Zhaoxiang Feng, Yulun Wu, Haoran Yuan, Yujie Zhao, Yu-Yang Qian, Bojun Wang, Peng Zhao, Daxin Jiang, Yibo Zhu, Tajana Rosing, Hao Zhang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 26&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Speculative decoding (SD) accelerates autoregressive Large Language Models (LLMs) by drafting multiple tokens and verifying them in parallel, but it faces a scaling limitation: increasing the draft budget improves speed only when acceptance remains high and drafting overhead stays low. This ceiling has been difficult to break because prior head-based SD methods face a causality-efficiency dilemma. Autoregressive drafters produce path-conditioned candidates that are effective for tree speculative decoding with higher acceptance length, but their drafting cost grows with tree depth. Bidirectional block-diffusion drafters generate all positions in one pass, but their branch-agnostic marginals can form individually plausible yet mutually inconsistent trees, wasting budget and reducing acceptance. We propose JetSpec, a head-based SD framework that combines one-forward drafting efficiency with branch-wise causal conditioning. JetSpec trains a causal parallel draft head over fused hidden states from the frozen target model, producing candidate trees whose scores align with the target model's autoregressive factorization. This enables JetSpec to convert larger draft budgets into longer accepted prefixes and higher end-to-end speedup. Across math, coding, and chat benchmarks on dense and MoE Qwen3 models, JetSpec consistently outperforms bidirectional-head and tree-based SD baselines. On H100 GPUs, JetSpec achieves up to 9.64x speedup on MATH-500 and 4.58x on open-ended conversational workloads, with further latency gains demonstrated through vLLM integration under realistic serving loads. Our code and models are available at https://github.com/hao-ai-lab/JetSpec.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20529</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.18394</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>ENPIRE: Agentic Robot Policy Self-Improvement in the Real World</title>
+      <title>Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation</title>
-      <link>https://arxiv.org/abs/2606.19980</link>
+      <link>https://arxiv.org/abs/2606.26907</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19980.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Wenli Xiao, Jia Xie, Tonghe Zhang, Haotian Lin, Letian "Max" Fu, Haoru Xue, Jalen Lu, Yi Yang, Cunxi Dai, Zi Wang, Jimmy Wu, Guanzhi Wang, S. Shankar Sastry, Ken Goldberg, Linxi "Jim" Fan, Yuke Zhu, Guanya Shi&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 10&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Achieving dexterous robotic manipulation in the real world heavily relies on human supervision and algorithm engineering, which becomes a central bottleneck in the pursuit of general physical intelligence. Although emerging coding agents can generate code to automate algorithm search, their successes remain largely confined in digital environments. We conjecture that the missing abstraction to automate robotics research is a repeatable feedback loop for real-world policy improvement: reset the scene, execute a policy, verify the outcome, and refine the next iteration. To bridge this gap, we introduce ENPIRE, a harness framework for coding agents that instantiates this physical feedback routine with four core modules: an Environment module (EN) for automatic reset and verification, a Policy Improvement module (PI) that launches policy refinement, a Rollout module (R) to evaluate policies with one or multiple physical robots operating in parallel, and an Evolution module (E) in which coding agents analyze logs, consult literature, improve training infrastructure and algorithm code to address failure modes. This closed-loop system transforms real-world manipulation learning into a controllable optimization procedure, minimizing human effort while allowing fair ablations across training recipe and agent variants. Powered by ENPIRE, frontier coding agents can autonomously train a policy to achieve a 99% success rate on challenging, dexterous manipulation tasks, such as organizing a pin box, fastening a zip tie, and tool use, a process that further accelerates when we dispatch an agent team on a robot fleet. Our results suggest a practical and scalable path toward deploying coding agents to autonomously advancing robotics in the physical world.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26907.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Zekai Zhang, Jiahao Li, Jie Zhang, Kaiyuan Gao, Kun Yan, Lihan Jiang, Ningyuan Tang, Shengming Yin, Tianhe Wu, Xiaoyue Chen, Xiao Xu, Yan Shu, Yanran Zhang, Yixian Xu, Yuxiang Chen, Zhendong Wang, Zihao Liu, Zikai Zhou, Huishuai Zhang, Dongyan Zhao, Chenfei Wu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 36&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; While text-to-image (T2I) models have achieved remarkable progress, they struggle with real-world requests that are often underspecified, implicit, or dependent on up-to-date knowledge. We identify this challenge as the Context Gap: the mismatch between the user context and the sufficient generation context for T2I models. To bridge this gap, we propose Qwen-Image-Agent, a unified agentic framework that integrates plan, reason, search, memory and feedback in a context-centric manner. Qwen-Image-Agent treats user input as partial context and progressively constructs the generation context through Context-Aware Planning and Context Grounding. Specifically, Context-Aware Planning identifies missing context and plans how it should be acquired and used, while Context Grounding gathers this context from reason, search, memory, and feedback. To evaluate agentic image generation, we further introduce Image Agent Bench (IA-Bench), a benchmark covering four core image agent capabilities: Plan, Reason, Search, and Memory. Experiments on IA-Bench, Mindbench and WISE-Verified show that Qwen-Image-Agent outperforms strong baselines and achieves state-of-the-art performance.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19980</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26907</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>Current World Models Lack a Persistent State Core</title>
+      <title>ViQ: Text-Aligned Visual Quantized Representations at Any Resolution</title>
-      <link>https://arxiv.org/abs/2606.20545</link>
+      <link>https://arxiv.org/abs/2606.27313</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20545.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Jinpeng Lu, Dexu Zhu, Haoyuan Shi, Linghan Cai, Guo Tang, Yinda Chen, Jie Cao, Duyu Tang, Yi Zhang, Yong Dai, Xiaozhu Ju&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 11&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; World models are increasingly regarded as a decisive step toward artificial general intelligence, yet modeling the physical world demands more than rendering convincing frames on demand: it requires an internal world state that keeps evolving over time, decoupled from observation, so that objects endure and events run to their conclusions whether or not a camera is watching, much as the moon holds to its orbit when no one is looking. This requirement is a blind spot of existing benchmarks, which reward surface properties such as fidelity, motion, and camera controllability while never asking whether a generated world keeps evolving once it is unobserved. We introduce WRBench, the first systematic diagnostic benchmark that treats camera motion as an intervention on observability and resolves evaluation into a human-calibrated chain that asks whether the camera executes the requested interaction, whether the scene stays continuous and identifiable while in view, and whether a returning target remains consistent with the event that was set in motion. Across 9{,}600 videos from 23 models spanning four control paradigms, one finding proves stubborn: current systems maintain the observed world as a tracking shot, resuming a returning target in the state at which it was abandoned rather than advancing the event while it went unseen. Because this failure recurs across control paradigms, model families, and increments of scale, robust world-state evolution does not follow from cleaner imagery, tighter control, richer geometric priors, or sheer parameter count We therefore argue that the stability of the physical state kernel and the consistency of worldlines under viewpoint intervention should become first-class objectives of world-model design, so that a world model captures how the world will unfold rather than how the next frame appears.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27313.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Xumin Yu, Zuyan Liu, Zhenyu Yang, Yuhao Dong, Shengsheng Qian, Jiwen Lu, Han Hu, Yongming Rao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 37&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same way as text inevitably introduces severe information loss. Existing work struggles to balance low-level details and high-level semantics in discrete representations: reconstruction-oriented representations often lack semantic information, whereas semantically stronger features typically suffer from severe loss of detail. We present ViQ, a Visual Quantized Representations framework, which is designed to balance semantics and details in discrete representations while supporting inputs at native resolutions, thereby enabling it to serve as a unified and general discrete representation for arbitrary visual inputs. Our approach structures quantization learning into two stages: text-aligned pre-training and feature discretization. With text-aligned pre-training, we enhance the visual encoder semantic-rich supervision from the pretrained language model and enable it to process native-resolution visual inputs. During discretization, we propose a proximal representation learning strategy to progressively compact the feature space, along with a position-aware head-wise quantization mechanism that enables flexible processing of arbitrary resolutions. Extensive experiments on multimodal tasks demonstrate that ViQ achieves competitive performance compared to state-of-the-art multimodal vision encoders with continuous and high-dimensional visual features, while maintaining high precision in low-level reconstruction. We also show that multimodal training with visual quantized representations largely improves efficiency, yielding up to 20\%-70\% acceleration with different base LLMs and training recipes.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20545</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27313</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>JanusMesh: Fast and Zero-Shot 3D Visual Illusion Generation via Cross-Space Denoising</title>
+      <title>OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning</title>
-      <link>https://arxiv.org/abs/2606.20563</link>
+      <link>https://arxiv.org/abs/2606.26790</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20563.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Siang-Ling Zhang, Huai-Hsun Cheng, Tsung-Ju Yang, Yu-Lun Liu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 19&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Creating 3D visual illusions, a single 3D mesh that reveals entirely different semantics from various viewing angles, is a fascinating but tough challenge. Existing optimization-based methods are slow and can produce oversaturated colors. In contrast, naive stitching approaches fail to produce geometrically coherent objects. This results in visible unnatural seams and semantic leaks. In this paper, we present a fast and training-free framework for generating text-driven 3D visual illusions. Our approach decouples the generation into two stages. First, we propose a cross-space dual-branch denoising process. This process dynamically decodes 3D latents into voxel space for CLIP-guided orientation alignment and Signed Distance Field (SDF) blending, which ensures seamless geometric fusion. Second, we introduce a view-conditioned texture synthesis module that projects and aggregates view-specific 2D diffusion priors onto the fused geometry. Extensive experiments demonstrate that our method generates highly realistic, dual-semantic 3D illusions in just 3-5 minutes. It significantly outperforms existing methods in geometric integrity, semantic recognizability, and efficiency. Project page: https://siang1105.github.io/JanusMesh.github.io/&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26790.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Shuo Yang, Jinyang Wu, Zhengxi Lu, Yuhao Shen, Fan Zhang, Lang Feng, Shuai Zhang, Haoran Luo, Zheng Lian, Zhengqi Wen, Jianhua Tao&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 38&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Outcome-based reinforcement learning provides a stable optimization backbone for language agents, but its sparse trajectory-level rewards provide little guidance on which intermediate decisions should be reinforced or suppressed. On-policy self-distillation offers dense token-level supervision, yet existing skill-conditioned variants often rely on external skill memories or retrieved privileged context, which are costly to maintain and can be mismatched with the state distribution induced by the current policy in multi-turn interaction. We propose OPID (On-Policy Skill Distillation), a framework that extracts skill supervision directly from completed on-policy trajectories. OPID represents trajectory hindsight as hierarchical skills: episode-level skills capture global workflows or failure-avoidance rules, while step-level skills capture local decision knowledge at critical timesteps. A critical-first routing mechanism uses step-level skills when critical decisions are identified and falls back to episode-level skills as default guidance otherwise. The selected skill is injected into the interaction history, allowing the old policy to re-score the same sampled response under both original and skill-augmented contexts. The resulting log-probability shift yields a token-level self-distillation advantage, which is combined with the outcome advantage for policy optimization. OPID thus preserves RL as the primary training objective while introducing dense, distribution-matched hindsight supervision. Experiments on ALFWorld, WebShop and Search-based QA demonstrate that OPID generally improves agent performance, sample efficiency, and robustness over outcome-only RL and existing skill-distillation baselines. Our code is available at https://github.com/jinyangwu/OPID/tree/main.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20563</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26790</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>FlowBender: Feedback-Aware Training for Self-Correcting Conditional Flows</title>
+      <title>In-Context World Modeling for Robotic Control</title>
-      <link>https://arxiv.org/abs/2606.20404</link>
+      <link>https://arxiv.org/abs/2606.26025</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20404.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Daniel Gilo, Sven Elflein, Ido Sobol, Or Litany&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 19&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Conditional diffusion and flow models routinely fail to satisfy the very constraints that define their task. For instance, a depth-conditioned model often produces images whose re-extracted depth disagrees with the input, even though the forward operator--the depth predictor defining the constraint--is available during both training and inference. Existing approaches generally fall into two categories: supervised models that treat the conditioning signal as a static cue and ignore alignment information at inference, and guidance-based methods that consult it through hand-tuned linear updates, typically trading fidelity to the condition against the plausibility of the generated sample. We argue that the fundamental gap in both paradigms is that the model is never trained to utilize its own alignment error. We introduce FlowBender, a closed-loop framework that treats this error as a first-class input, training the network to learn a correction policy conditioned on inference-time feedback. At each step, an unguided look-ahead pass estimates the clean signal, a task-specific deviation is computed via the forward operator, and a refinement pass consumes this signal to produce a corrected velocity. We propose several variants of FlowBender, including a gradient-based formulation for differentiable operators and a zero-order variant for non-differentiable settings such as JPEG compression. For efficient sampling, we introduce a prior-step shortcut that enables closed-loop correction at a minimal additional computational cost. Across image-to-image translation, restoration, and 3D mesh texturing, FlowBender consistently outperforms standard supervised baselines, alignment-loss-augmented training, and state-of-the-art inference-time guidance, improving fidelity and plausibility simultaneously rather than trading them against each other. Project page: https://flow-bender.github.io/&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.26025.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Siyin Wang, Junhao Shi, Senyu Fei, Zhaoyang Fu, Li Ji, Jingjing Gong, Xipeng Qiu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 39&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Modern Vision-Language-Action (VLA) models often fail to generalize to novel setups, such as altered camera viewpoints or robot morphologies, because they are typically conditioned only on current observations and language instructions. By ignoring the underlying system configuration as a variable, these models implicitly assume a fixed execution context encountered during training, necessitating data-intensive fine-tuning for any new environment. In this work, we introduce In-Context World Modeling (ICWM), a framework that treats system identification as an in-context adaptation problem. ICWM enables robot policies to autonomously infer essential system variables from a short history of self-generated, task-agnostic interactions. Unlike traditional In-Context Learning that uses demonstrations to specify what task to perform, ICWM leverages the context window to understand how the system operates. By processing these interactions before task execution, the model implicitly captures the world dynamics of the current system, enabling adaptation to novel configurations without parameter updates. Extensive experiments in simulation and on real-world robot platforms demonstrate that ICWM significantly outperforms standard VLA baselines on novel camera viewpoints.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20404</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.26025</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
     </item>
     <item>
-      <title>FreeStyle: Free Control of Style-Content Dual-Reference Generation from Community LoRA Mining</title>
+      <title>DanceOPD: On-Policy Generative Field Distillation</title>
-      <link>https://arxiv.org/abs/2606.20506</link>
+      <link>https://arxiv.org/abs/2606.27377</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20506.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Jinghong Lan, Wei Cheng, Yunuo Chen, Ziqi Ye, Peng Xing, Yixiao Fang, Rui Wang, Yufeng Yang, Xuanyang Zhang, Xianfang Zeng, Difan Zou, Gang Yu, Chi Zhang&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 25&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Style-content dual-reference generation aims to synthesize an image that preserves the structure and semantics of a content reference while adopting the style of a separate style reference.Despite recent progress, this setting remains challenging because models must balance content fidelity, style alignment, and instruction following avoiding semantic leakage from the style reference.A key bottleneck is the lack of large-scale triplet data with clean content-style separation and broad long-tail style coverage.In this work, we propose FreeStyle, a scalable dual-reference generation framework based on community LoRA mining.We treat community LoRAs as compositional anchors for style and content, and design a rigorous generation and filtering pipeline to construct large-scale Style-Reference and Content-Reference triplets across multiple base models.To address content leakage, we adopt a two-stage curriculum with stage-specific disentanglement mechanisms: an attention-level enrichment constraint that suppresses style-reference leakage in the style-transfer stage, and a frequency-aware RoPE modulation strategy that targets positional-correspondence-based leakage in the harder dual-reference stage.We also introduce a benchmark covering both style-reference and dual-reference generation, with evaluations on style similarity, content preservation, aesthetics, instruction following, and leakage rejection. The benchmark incorporates a style-invariant Content Alignment Score (CAS) and introduces a calibrated VLM-based Rejection Score for evaluating generation reliability and leakage suppression.Extensive experiments show that our model achieves a strong balance among style alignment, content preservation, and leakage suppression.&lt;/p&gt;</description>
+      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.27377.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Wei Zhou, Xiongwei Zhu, Zelin Xu, Bo Dong, Lixue Gong, Yongyuan Liang, Meng Chu, Leigang Qu, Lingdong Kong, Wei Liu, Tat-Seng Chua&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 57&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Modern image generation demands a single model that unifies diverse capabilities, including text-to-image (T2I), local editing, and global editing. However, these capabilities are rarely naturally aligned and often conflict. For instance, editing tends to degrade T2I performance, while global and local editing interfere with each other. Consequently, effectively composing these capabilities has become a central challenge for image generation model training. To tackle this, we introduce DanceOPD, an on-policy generative field distillation framework for flow-matching models that routes each sample to one capability field, queries one low-noise student-induced state, and trains with a simple velocity MSE objective. With each capability source defined as a velocity field over the shared flow state space, the student learns from fields queried on its own rollout states to compose expert capabilities. This formulation also absorbs operator-defined fields such as classifier-free guidance. Comprehensive experiments on T2I, editing, realism-field absorption, and CFG absorption show that our approach improves multi-capability composition, strengthening target capabilities while preserving anchor generation quality. We believe this work establishes a practical route for generative field distillation in flow-matching models.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20506</guid>
+      <guid isPermaLink="false">https://arxiv.org/abs/2606.27377</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
+      <pubDate>Thu, 25 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>Beyond Static Leaderboards: Predictive Validity for the Evaluation of LLM Agents</title>
-      <link>https://arxiv.org/abs/2606.19704</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.19704.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Dhaval C. Patel, Kaoutar El Maghraoui, Shuxin Lin, Yusheng Li, Tianjun Feng, Chun-Yi Tsai, Yihan Sun, Wei Alexander Xin, Akshat Bhandari, Tanisha Rathod, Aaron Fan, Sanskruti Vijay Shejwal, Tomas Pasiecznik, Sagar Chethan Kumar, Tanmay Agarwal, Rohith Kanathur, Sam Colman, Amaan Sheikh, Dev Bahl, Ann Li, Krish Veera, Alimurtaza Mustafa Merchant, Shambhawi Baswaraj Bhure, Sajal Kumar Goyla, Chengrui Li, Kirthana Natarajan, Rui Li, Thomas Ajai, Rujing Li, Vivek G. Iyer, Sanjaii Vijayakumar, Yitong Bai, Ayal Yakobe, Darief Maes, Yassine Jebbouri, Tianyang Xu, Thai Quoc On, Vera Mazeeva, Winston Li, Yuval Shemla, Yeshitha Bhuvanesh, Rushin Bhatt, Siddharth Chethan Gowda, Alisha Vinod, Caroline Cahill, Shriya Aishani Rachakonda, Yunfeng Chen, Aryaman Agrawal, Aman Upganlawar, Mao Le Jonathan Ang, Yubin Sally Go, Madhav Rajkondawar, Yang-Jung Chen, Trisha Maturi, Ananya Kapoor, Andrew Li, Shrey Arora, Mana Abbaszadeh, Shen Li, Charles Xu, Byeolah Kwon&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 29&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Agent benchmarks are growing fast, but no single benchmark touches more than four or five of the dimensions that deployment exposes. This paper aggregates the largest coordinated deep-dive of one MCP-based industrial-agent benchmark to date: fourteen parallel implementation studies covering new asset classes (including a multi-modal visual extension), alternative orchestrations, retrieval strategies, reasoning modes, infrastructure optimizations, and evaluation-methodology probes. Consolidating those studies with seven prior agent benchmarks, we argue that aggregate-score leaderboards systematically underspecify deployed-agent evaluation. Rankings derived from aggregate scores do not transfer to out-of-distribution settings; recent public-to-hidden competition retrospectives provide direct empirical evidence of this rank instability. We propose ranking configurations by predictive validity, the correlation between in-sample and out-of-sample rank, rather than in-sample mean, and report a twelve-tier measurement apparatus that exposes the deployment-relevant dimensions HELM and its agent-era successors collapse. The position is operationalized through three falsifiable out-of-distribution criteria with explicit thresholds; existing evidence partly supports it but is too thin to confirm. We close with a pre-registered pilot design and a field-level vision for what the next generation of agentic benchmarks should report.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.19704</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>DF3DV-1K: A Large-Scale Dataset and Benchmark for Distractor-Free Novel View Synthesis</title>
-      <link>https://arxiv.org/abs/2604.13416</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2604.13416.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Cheng-You Lu, Yi-Shan Hung, Wei-Ling Chi, Hao-Ping Wang, Charlie Li-Ting Tsai, Yu-Cheng Chang, Yu-Lun Liu, Thomas Do, Chin-Teng Lin&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 29&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Advances in radiance fields have enabled photorealistic novel view synthesis. In several domains, large-scale real-world datasets have been developed to support comprehensive benchmarking and to facilitate progress beyond scene-specific reconstruction. However, for distractor-free radiance fields, a large-scale dataset with clean and cluttered images per scene remains lacking, limiting the development. To address this gap, we introduce DF3DV-1K, a large-scale real-world dataset comprising 1,048 scenes, each providing clean and cluttered image sets for benchmarking. In total, the dataset contains 89,924 images captured using consumer cameras to mimic casual capture, spanning 128 distractor types and 161 scene themes across indoor and outdoor environments. A curated subset of 41 scenes, DF3DV-41, is systematically designed to evaluate the robustness of distractor-free radiance field methods under challenging scenarios. Using DF3DV-1K, we benchmark nine recent distractor-free radiance field methods and 3D Gaussian Splatting, identifying the most robust methods and the most challenging scenarios. Beyond benchmarking, we demonstrate an application of DF3DV-1K by fine-tuning a diffusion-based 2D enhancer to improve radiance field methods, achieving average improvements of 0.96 dB PSNR and 0.057 LPIPS on the held-out set (e.g., DF3DV-41) and the On-the-go dataset. We hope DF3DV-1K facilitates the development of distractor-free vision and promotes progress beyond scene-specific approaches. The dataset and leaderboard are available at https://johnnylu305.github.io/df3dv1k_web/.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2604.13416</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>S-Agent: Spatial Tool-Use Elicits Reasoning for Spatial Intelligence</title>
-      <link>https://arxiv.org/abs/2606.20515</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20515.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Yalun Dai, Hao Li, Shulin Tian, Runmao Yao, Yuhao Dong, Fangzhou Hong, Zhaoxi Chen, Fangfu Liu, Baoliang Tian, Dingwen Zhang, Tao Wang, Kim-Hui Yap, Ziwei Liu&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 35&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; Real-world spatial intelligence requires reasoning over a continuous and evolving 3D world, yet existing VLMs and tool-augmented agents largely remain tied to static, stateless inference from isolated visual observations. We introduce \textsc{S-Agent}, a spatial tool-use agentic paradigm for understanding and reasoning over continuous multi-view images and videos. By formulating spatial reasoning as spatio-temporal evidence accumulation rather than isolated frame-level prediction, S-Agent reshapes spatial perception into scene-centric understanding beyond frame-centric recognition. Specifically, S-Agent casts the VLM as a semantic planner that decides what evidence is needed, while a hierarchy of spatial tools and experts grounds objects in 2D, lifts them into 3D geometric evidence, and aggregates this evidence into high-level spatial knowledge (e.g., counting, measurement, orientation, and relative position). Additionally, a temporal memory mechanism, including Scene Memory for maintaining the evolving scene state and Agent Memory for accumulating reasoning context, enables evidence integration across frames and reasoning steps. Comprehensive experiments on multi-view and video spatial reasoning benchmarks show that S-Agent consistently improves both open-source and closed-source VLMs in a training-free manner. Beyond inference-time augmentation, supervised fine-tuning (SFT) on S-Agent-generated spatial trajectories S-300K yields S-Agent-8B, a compact spatial agent that significantly surpasses similar-scale baselines (e.g., Qwen3-VL-8B) and performs comparably to advanced closed-source models (e.g., GPT-5.4 and Gemini 3).&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20515</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
-    </item>
-    <item>
-      <title>Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages</title>
-      <link>https://arxiv.org/abs/2606.20517</link>
-      <description>&lt;p&gt;&lt;img src="https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2606.20517.png" alt="Paper thumbnail" style="max-width: 300px; height: auto;" /&gt;&lt;/p&gt;&lt;p&gt;&lt;b&gt;Authors:&lt;/b&gt; Maria Ivanova, Pavel Zadorozhny, Rodion Levichev, Ivan Petrov, Adamenko Pavel, Ivan Lopatin, Alexey Kutalev, Dmitrii Babaev&lt;/p&gt;&lt;p&gt;&lt;b&gt;Upvotes:&lt;/b&gt; 49&lt;/p&gt;&lt;p&gt;&lt;b&gt;Summary:&lt;/b&gt; LiveCodeBench (LCB) has recently become a widely adopted benchmark for evaluating large language models (LLMs) on code-generation tasks. By curating competitive programming problems, constantly adding fresh problems to the set, and filtering them by release dates, LCB provides contamination-aware evaluation and offers a holistic view of coding capability. However, LCB remains restricted to Python, leaving open the question of whether LLMs can generalize across the diverse programming languages required in real-world software engineering.   We introduce Multi-LCB, a benchmark for evaluating LLMs across twelve programming languages, including Python. Multi-LCB transforms Python tasks from the LCB dataset into equivalent tasks in other languages while preserving LCB's contamination controls and evaluation protocol. Because it is fully compatible with the original LCB format, Multi-LCB will automatically track future LCB updates, enabling systematic assessment of cross-language code generation competence and requiring models to sustain performance well beyond Python.   We evaluated 24 LLMs for instruction and reasoning on Multi-LCB, uncovering evidence of Python overfitting, language-specific contamination, and substantial disparities in multilingual performance. Our results establish Multi-LCB as a rigorous new benchmark for multi-programming-language code evaluation, directly addressing LCB's primary limitation and exposing critical gaps in current LLM capabilities.&lt;/p&gt;</description>
-      <guid isPermaLink="false">https://arxiv.org/abs/2606.20517</guid>
-      <pubDate>Thu, 18 Jun 2026 00:00:00 +0000</pubDate>
     </item>
   </channel>
 </rss>

run.py

@@ -0,0 +1,92 @@
+#!/usr/bin/env python3
+"""
+run.py — entrypoint for HF Daily Paper RSS generation.
+
+Generates:
+  - Daily feed (always)
+  - Weekly feed (every Monday)
+  - Monthly feed (1st of month)
+
+Usage:
+  python run.py                              # output to ./feeds/
+  python run.py --output-dir /data/feeds
+  python run.py --date 2026-06-27            # simulate date
+"""
+import argparse
+import datetime
+import os
+import subprocess
+import sys
+
+
+def main():
+    parser = argparse.ArgumentParser(description="HF Daily Paper RSS Generator")
+    parser.add_argument("--output-dir", default=os.path.join(os.getcwd(), "feeds"),
+                        help="Directory to save RSS feeds")
+    parser.add_argument("--date", default=None,
+                        help="Simulate a specific date (YYYY-MM-DD)")
+    args = parser.parse_args()
+
+    output_dir = os.path.abspath(args.output_dir)
+    os.makedirs(output_dir, exist_ok=True)
+
+    # Resolve date
+    if args.date:
+        today = datetime.date.fromisoformat(args.date)
+    else:
+        today = datetime.date.today()
+
+    year = today.year
+    month = f"{today.month:02d}"
+    week_num = today.isocalendar()[1]
+    is_monday = today.isoweekday() == 1
+    is_first = today.day == 1
+
+    script_dir = os.path.dirname(os.path.abspath(__file__))
+    parser_py = os.path.join(script_dir, "parser.py")
+
+    print(f"[run.py] Date: {today.isoformat()} "
+          f"(Year={year} Month={month} Week={week_num} "
+          f"Mon={is_monday} First={is_first})")
+    print(f"[run.py] Output: {output_dir}")
+
+    def run_parser(source, outfile):
+        full_path = os.path.join(output_dir, outfile)
+        print(f"\n── Generating {outfile} ──")
+        result = subprocess.run(
+            [sys.executable, parser_py, "--source", source, "--output", full_path],
+            capture_output=False, text=True, cwd=script_dir,
+        )
+        return result.returncode
+
+    # 1. Daily (always)
+    run_parser("https://huggingface.co/papers", "feed.xml")
+
+    # 2. Weekly (Mondays)
+    if is_monday:
+        run_parser(
+            f"https://huggingface.co/papers/week/{year}-W{week_num}",
+            "feed_weekly.xml",
+        )
+    else:
+        print("[run.py] Skipping weekly (not Monday)")
+
+    # 3. Monthly (1st)
+    if is_first:
+        run_parser(
+            f"https://huggingface.co/papers/month/{year}-{month}",
+            "feed_monthly.xml",
+        )
+    else:
+        print("[run.py] Skipping monthly (not 1st)")
+
+    print(f"\n[run.py] ✅ Done — feeds in {output_dir}")
+    for f in sorted(os.listdir(output_dir)):
+        if f.endswith(".xml"):
+            fpath = os.path.join(output_dir, f)
+            size = os.path.getsize(fpath)
+            print(f"  {f}  ({size:,} bytes)")
+
+
+if __name__ == "__main__":
+    main()

run.sh

@@ -0,0 +1,16 @@
+#!/bin/bash
+set -euo pipefail
+# ─────────────────────────────────────────────
+# HuggingFace Daily Paper RSS — Entrypoint Script
+# ─────────────────────────────────────────────
+# Convenience wrapper — delegates to run.py
+#
+# Usage:
+#   ./run.sh                          # output to ./feeds/
+#   ./run.sh --output-dir /data/feeds
+#   ./run.sh --date 2026-06-27        # simulate date
+# ─────────────────────────────────────────────
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PYTHON=$(command -v python3 || command -v python)
+exec "$PYTHON" "$SCRIPT_DIR/run.py" "$@"

server.py

@@ -0,0 +1,139 @@
+#!/usr/bin/env python3
+"""
+RSS Feed HTTP Server — generates feeds on startup + serves them via HTTP.
+For EasyPanel deployment: single entrypoint that handles everything.
+
+Usage:
+    python server.py                         # port 8080, dir /data
+    python server.py --dir /data --port 8080
+    python server.py --dir /data --port 8080 --refresh 24
+"""
+import argparse
+import datetime
+import http.server
+import logging
+import os
+import subprocess
+import sys
+import threading
+import time
+
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s [%(levelname)s] %(message)s",
+    datefmt="%Y-%m-%d %H:%M:%S",
+)
+logger = logging.getLogger("hf-rss")
+
+
+# ── Feed generation ──
+
+def generate_feeds(output_dir: str):
+    """Run run.py to generate RSS feeds."""
+    script_dir = os.path.dirname(os.path.abspath(__file__))
+    run_py = os.path.join(script_dir, "run.py")
+    python = sys.executable
+
+    logger.info("Generating RSS feeds...")
+    try:
+        result = subprocess.run(
+            [python, run_py, "--output-dir", output_dir],
+            capture_output=True, text=True, timeout=60,
+            cwd=script_dir,
+        )
+        if result.returncode == 0:
+            logger.info(f"Feeds generated successfully in {output_dir}")
+            for f in sorted(os.listdir(output_dir)):
+                if f.endswith(".xml"):
+                    fpath = os.path.join(output_dir, f)
+                    size = os.path.getsize(fpath)
+                    logger.info(f"  {f}  ({size:,} bytes)")
+        else:
+            logger.error(f"Feed generation failed (exit {result.returncode})")
+            logger.error(result.stderr[:500] if result.stderr else result.stdout[:500])
+    except Exception as e:
+        logger.error(f"Feed generation error: {e}")
+
+
+def background_refresh(output_dir: str, interval_hours: int):
+    """Regenerate feeds periodically."""
+    if interval_hours <= 0:
+        return
+    interval_seconds = interval_hours * 3600
+    logger.info(f"Background refresh every {interval_hours}h (next in {interval_hours}h)")
+    while True:
+        time.sleep(interval_seconds)
+        generate_feeds(output_dir)
+
+
+# ── HTTP Server ──
+
+class RSSFeedHandler(http.server.SimpleHTTPRequestHandler):
+    """Custom handler with CORS headers."""
+
+    def __init__(self, *args, directory=None, **kwargs):
+        super().__init__(*args, directory=directory, **kwargs)
+
+    def end_headers(self):
+        self.send_header("Access-Control-Allow-Origin", "*")
+        self.send_header("Access-Control-Allow-Methods", "GET, OPTIONS")
+        self.send_header("Cache-Control", "no-cache, no-store, must-revalidate")
+        self.send_header("Pragma", "no-cache")
+        self.send_header("Expires", "0")
+        super().end_headers()
+
+    def do_OPTIONS(self):
+        self.send_response(204)
+        self.end_headers()
+
+
+# ── Main ──
+
+def main():
+    parser = argparse.ArgumentParser(description="HF Daily Paper RSS HTTP Server")
+    parser.add_argument("--dir", default="/data",
+                        help="Feed output directory (default: /data)")
+    parser.add_argument("--port", type=int, default=8080,
+                        help="HTTP server port (default: 8080)")
+    parser.add_argument("--host", default="0.0.0.0",
+                        help="Bind address (default: 0.0.0.0)")
+    parser.add_argument("--refresh", type=int, default=24,
+                        help="Auto-refresh interval in hours (0=disabled, default: 24)")
+    args = parser.parse_args()
+
+    feed_dir = os.path.abspath(args.dir)
+    os.makedirs(feed_dir, exist_ok=True)
+
+    # Generate feeds on startup
+    generate_feeds(feed_dir)
+
+    # Start background refresh thread
+    if args.refresh > 0:
+        t = threading.Thread(
+            target=background_refresh,
+            args=(feed_dir, args.refresh),
+            daemon=True,
+        )
+        t.start()
+
+    # Start HTTP server
+    class Handler(RSSFeedHandler):
+        def __init__(self, *args, **kwargs):
+            super().__init__(*args, directory=feed_dir, **kwargs)
+
+    server = http.server.HTTPServer((args.host, args.port), Handler)
+    logger.info(f"Serving RSS feeds from {feed_dir}")
+    logger.info(f"Listening on http://{args.host}:{args.port}")
+    logger.info(f"  GET /feed.xml           (daily)")
+    logger.info(f"  GET /feed_weekly.xml    (weekly, Mondays)")
+    logger.info(f"  GET /feed_monthly.xml   (monthly, 1st)")
+
+    try:
+        server.serve_forever()
+    except KeyboardInterrupt:
+        logger.info("Shutting down...")
+        server.server_close()
+
+
+if __name__ == "__main__":
+    main()