Daily Papers of 2026-10-15

  1. Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation 367 upvotes, #1 of 2026-10-15
  2. Player-1: Learning to Play from Synthetic Experience 349 upvotes, #2 of 2026-10-15
  3. Beyond Domain-Routed Supervision: Cross-Teacher Guarding for Multi-Teacher On-Policy Distillation 94 upvotes, #3 of 2026-10-15
  4. Speculative Sampling in Interactive Video Generation 84 upvotes, #4 of 2026-10-15
  5. LatentLoop: Scaling Looped MoE Models with Asymmetric Compute Allocation 81 upvotes, #5 of 2026-10-15
  6. Many Right Answers, One Training Example: Removing Responder Noise from GRPO for LLM Recommenders 79 upvotes, #6 of 2026-10-15
  7. Cortex: Consolidating Context into Parametric Memory 73 upvotes, #7 of 2026-10-15
  8. ActionOPD: Distilling Decisions, Not Tokens 64 upvotes, #8 of 2026-10-15
  9. Doc2Harness: Benchmarking LLMs on Long-Horizon Harness Engineering of Coding Agents 62 upvotes, #9 of 2026-10-15
  10. More Methods, Better Agents? A Reality Check on Self-Distillation in Agentic Reinforcement Learning 58 upvotes, #10 of 2026-10-15
  11. Seeing, Maintaining, and Learning: The Evolution of Multimodal Agent Memory 55 upvotes, #11 of 2026-10-15
  12. LensErase: A Foundation Model for Erase Anything 52 upvotes, #12 of 2026-10-15
  13. KV4-RL: Unlocking Faster Asynchronous Rollouts for RLVR with INT4 KV Cache Quantization 49 upvotes, #13 of 2026-10-15
  14. Recovering Reliable Supervision via Critique-Guided On-Policy Distillation 45 upvotes, #14 of 2026-10-15
  15. Towards Agentic Robots: Model, Data, Environment, and Harness 43 upvotes, #15 of 2026-10-15
  16. Hybrid Reasoning Agents 38 upvotes, #16 of 2026-10-15
  17. RuleTrek: Benchmarking Agentic Rule Reasoning in Complex Scenarios 38 upvotes, #16 of 2026-10-15
  18. HumanoidToolBench: Benchmarking Humanoid Tool Use from Selection to Mobile Execution 36 upvotes, #18 of 2026-10-15
  19. CLASHarness: Benchmarking Multi-Agent Conflict in Computer Use via a Controllable and Verifiable Generation Pipeline 36 upvotes, #18 of 2026-10-15
  20. SwarmRSI: Recursive Self-Improvement through Evolving Scientific Agent Swarm 34 upvotes, #20 of 2026-10-15
  21. The Thinking Leakage in Reinforcement Learning for LLMs 33 upvotes, #21 of 2026-10-15
  22. APT: Feedback-Driven Agentic Post-Training for Text-to-Image Generation 31 upvotes, #22 of 2026-10-15
  23. Long-Horizon RL Needs Muon 30 upvotes, #23 of 2026-10-15
  24. Apertus 1.5: From Open Text Models to Open Multimodal Reasoners 29 upvotes, #24 of 2026-10-15
  25. τ-Emotion-Bench: Exposing the Interaction Fragility of LLM Agents through Real-World User Emotions 28 upvotes, #25 of 2026-10-15
  26. APEX-Harm: Evaluating Agent Safety in Shared Professional Workspaces 26 upvotes, #26 of 2026-10-15
  27. Super-Resolution in The Right Latent Space: A Frozen Vision-Foundation Substrate 25 upvotes, #27 of 2026-10-15
  28. StartLux-Decision: AutoThink Decision Models with Reinforcement Learning for Calibrated Deliberative Decisions 25 upvotes, #27 of 2026-10-15
  29. DecisionRobust: Benchmarking Robustness for JEV Style Decision Models Across Text and Vision 24 upvotes, #29 of 2026-10-15
  30. Efficient Scene-Text Refinement via Pixel-Space Grouped Image Generation 22 upvotes, #30 of 2026-10-15
  31. RAR: Rate-Adaptive Rollout for Time-Consistent Interactive World Models 21 upvotes, #31 of 2026-10-15
  32. LabInstruct: Benchmarking Situated Instructional Video Generation for Lab Procedures 20 upvotes, #32 of 2026-10-15
  33. Share the Symbol, Not the Vector: Training-Free Fact Communication Across Language Models 20 upvotes, #32 of 2026-10-15
  34. ProgAgent: Learning Tool Orchestration through Programmatic Training 19 upvotes, #34 of 2026-10-15
  35. History Replays Itself: Unleashing Long-Context Capability in Linear Attention 19 upvotes, #34 of 2026-10-15
  36. Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models 19 upvotes, #34 of 2026-10-15
  37. A Diagnostic Study of World-Space Human Recovery Under High Camera Dynamics 19 upvotes, #34 of 2026-10-15
  38. SkillMarker: A KGW-Based Watermarking Method for Agent Skill Ownership Verification 18 upvotes, #38 of 2026-10-15
  39. GroundGUI: GUI Grounding by High-Quality Data and Distillation-to-Reinforcement Learning 17 upvotes, #39 of 2026-10-15
  40. What Makes a Social Agent's Interactions Effective? Learning from Verifiable Outcomes in Paired Worlds 17 upvotes, #39 of 2026-10-15
  41. Joint Subspace Sparse Attention for Efficient Long-Context Decoding 16 upvotes, #41 of 2026-10-15
  42. SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving 16 upvotes, #41 of 2026-10-15
  43. Next-Concept Modeling: Fast, Parallel Coarse-to-Fine Text Generation 15 upvotes, #43 of 2026-10-15
  44. FreeSolo: Linear-time Sparse Attention with Zero-Shot Length Extrapolation 15 upvotes, #43 of 2026-10-15
  45. LLM-to-Flow Distillation: Teaching Natural Language Goals to Scientific Generators 15 upvotes, #43 of 2026-10-15
  46. Self-Conditioning Enables Deep Sequential Reasoning in Flow Matching 15 upvotes, #43 of 2026-10-15
  47. BigCell: Generating Gigapixel Whole-Slide Images 13 upvotes, #47 of 2026-10-15
  48. ProbabilityBench: Model Evaluation and Calibration with Exact Ground-Truth Uncertainty 13 upvotes, #47 of 2026-10-15
  49. Diagnosing and Correcting Reward Hacking in Verifiable Instruction Following 13 upvotes, #47 of 2026-10-15
  50. Glance Forcing: Making Bidirectional Diffusion Autoregressive at a Glance 12 upvotes, #50 of 2026-10-15
  51. Rethinking Projector Training in Multimodal LLMs 11 upvotes, #51 of 2026-10-15
  52. From Failures to Learnable Worlds: Policy-Conditioned Environment Generation for Navigation Agents 10 upvotes, #52 of 2026-10-15
  53. The COUNT: Counting N-grams Unlocks Scaling on Host Memory and Training-Free Adaptation 10 upvotes, #52 of 2026-10-15
  54. Continuum Attention 10 upvotes, #52 of 2026-10-15
  55. VITAL: From Visual Motion to Precise Action 10 upvotes, #52 of 2026-10-15
  56. Acting Through Registers 9 upvotes, #56 of 2026-10-15
  57. PreciseCoder: Faithful and Generalizable Debugging Beyond Unit-Test Rewards 9 upvotes, #56 of 2026-10-15
  58. MA-Bench: Measuring Operational Planning in Language Models 9 upvotes, #56 of 2026-10-15
  59. TextVerifier: Seeing Beyond Recognition to Verify Chinese Text Rendering 9 upvotes, #56 of 2026-10-15
  60. VASPO: Self-Distillation with Privileged Value Estimation 9 upvotes, #56 of 2026-10-15
  61. Latent Compaction with Test-Time Training 9 upvotes, #56 of 2026-10-15
  62. SWEeper-Bench: Can Agents Discover Bugs in Interactive Software? 8 upvotes, #62 of 2026-10-15
  63. Exploring the Limits of Combinatorial Generalization in Embodied Manipulation 8 upvotes, #62 of 2026-10-15
  64. Rethinking LLM-Guided Visual Question Answering: Does Deeper Knowledge Integration Really Help? 8 upvotes, #62 of 2026-10-15
  65. Diff Diffusion: Predicting Edit Paths with Diffusion Models for Efficient Code Editing 8 upvotes, #62 of 2026-10-15
  66. Training LLMs for Multi-Agent Collaboration with Verifiable Information Rewards 8 upvotes, #62 of 2026-10-15
  67. Generative World Simulation with DAO and JING 8 upvotes, #62 of 2026-10-15
  68. SyncReward: Human-Aligned Synchronization Reward for Audio-Visual Generation 7 upvotes, #68 of 2026-10-15
  69. Choosing LLMs for Strategic Agent Interactions 7 upvotes, #68 of 2026-10-15
  70. Scaling Online Optimal Transport Minibatches for Flow Matching 7 upvotes, #68 of 2026-10-15
  71. CoTranslator: Reverse to Forward Chain-of-Thought Translator for Reasoning Distillation 7 upvotes, #68 of 2026-10-15
  72. MimicX: Policy-in-the-Loop Supervision Refinement for Video-Driven Humanoid Motion Tracking 6 upvotes, #72 of 2026-10-15
  73. NavGPT-3: Harnessing Context in a Hierarchical Navigation Runtime 6 upvotes, #72 of 2026-10-15
  74. Jag: Direct Box Prediction for Efficient Visual Grounding 6 upvotes, #72 of 2026-10-15
  75. When Left Is Right? Camera-Motion Understanding beyond Screen 6 upvotes, #72 of 2026-10-15
  76. To Align or Not To Align: Check Your COMPASS Before You Train 6 upvotes, #72 of 2026-10-15
  77. Unlocking Structured Pruning for Reasoning Language Models via Attention-Preserving MLP Trimming 6 upvotes, #72 of 2026-10-15
  78. AgentSafeEvo: Adversarial Co-Evolution for LLM Agent Safety 6 upvotes, #72 of 2026-10-15
  79. Physical Time as a Missing Dimension of Robot Policy Memory Dynamics 6 upvotes, #72 of 2026-10-15
  80. Your Transformer Is Secretly Elastic 6 upvotes, #72 of 2026-10-15
  81. G^3VLA: Geometric Inductive Bias for Vision-Language-Action Models 5 upvotes, #81 of 2026-10-15
  82. Vela: Scaling Vision-Language-Action Models with Adaptive Action Curve Parametrization 5 upvotes, #81 of 2026-10-15
  83. StreamErase: Towards Ultra-Fast Streaming Video Erasing 5 upvotes, #81 of 2026-10-15
  84. Cortex: Measurement-Gated Evolution and Pruning of Coding-Agent Skills and Rules 4 upvotes, #84 of 2026-10-15
  85. The Agent Knew: Internal States Forecast Unauthorized Actions and Stop Them, and the Lies That Follow 4 upvotes, #84 of 2026-10-15
  86. Looking Complete Is Not Being Grounded: Evaluating LLM Sitreps Against Human-Authored Situation Reports 3 upvotes, #86 of 2026-10-15
  87. VLX-Go: Robust Embodied Tracking through Multi-Stage Learning and On-Policy Refinement 3 upvotes, #86 of 2026-10-15
  88. Lluna: Collaborative Unimodal Models for Missing-Modality Inference 3 upvotes, #86 of 2026-10-15
  89. Drive3R: Block-wise Streaming Geometric Foundation Models for Autonomous Driving 1 upvotes, #89 of 2026-10-15

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.