Daily Papers of 2026-05-08

  1. Skill1: Unified Evolution of Skill-Augmented Agents via Reinforcement Learning 106 upvotes, #1 of 2026-05-08
  2. Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction 105 upvotes, #2 of 2026-05-08
  3. Continuous Latent Diffusion Language Model 75 upvotes, #3 of 2026-05-08
  4. MiA-Signature: Approximating Global Activation for Long-Context Understanding 54 upvotes, #4 of 2026-05-08
  5. RaguTeam at SemEval-2026 Task 8: Meno and Friends in a Judge-Orchestrated LLM Ensemble for Faithful Multi-Turn Response Generation 44 upvotes, #5 of 2026-05-08
  6. When to Trust Imagination: Adaptive Action Execution for World Action Models 42 upvotes, #6 of 2026-05-08
  7. SkillOS: Learning Skill Curation for Self-Evolving Agents 42 upvotes, #6 of 2026-05-08
  8. MARBLE: Multi-Aspect Reward Balance for Diffusion RL 39 upvotes, #8 of 2026-05-08
  9. Nonsense Helps: Prompt Space Perturbation Broadens Reasoning Exploration 36 upvotes, #9 of 2026-05-08
  10. Audio-Visual Intelligence in Large Foundation Models 32 upvotes, #10 of 2026-05-08
  11. Continuous-Time Distribution Matching for Few-Step Diffusion Distillation 25 upvotes, #11 of 2026-05-08
  12. StraTA: Incentivizing Agentic Reinforcement Learning with Strategic Trajectory Abstraction 25 upvotes, #11 of 2026-05-08
  13. Auto Research with Specialist Agents Develops Effective and Non-Trivial Training Recipes 15 upvotes, #13 of 2026-05-08
  14. AI Co-Mathematician: Accelerating Mathematicians with Agentic AI 15 upvotes, #13 of 2026-05-08
  15. A^2TGPO: Agentic Turn-Group Policy Optimization with Adaptive Turn-level Clipping 14 upvotes, #15 of 2026-05-08
  16. Can RL Teach Long-Horizon Reasoning to LLMs? Expressiveness Is Key 14 upvotes, #15 of 2026-05-08
  17. UniPool: A Globally Shared Expert Pool for Mixture-of-Experts 11 upvotes, #17 of 2026-05-08
  18. EMO: Pretraining Mixture of Experts for Emergent Modularity 10 upvotes, #18 of 2026-05-08
  19. ReflectDrive-2: Reinforcement-Learning-Aligned Self-Editing for Discrete Diffusion Driving 9 upvotes, #19 of 2026-05-08
  20. RemoteZero: Geospatial Reasoning with Zero Human Annotations 8 upvotes, #20 of 2026-05-08
  21. TabEmbed: Benchmarking and Learning Generalist Embeddings for Tabular Understanding 8 upvotes, #20 of 2026-05-08
  22. TIDE: Every Layer Knows the Token Beneath the Context 8 upvotes, #20 of 2026-05-08
  23. Balanced Aggregation: Understanding and Fixing Aggregation Bias in GRPO 7 upvotes, #23 of 2026-05-08
  24. KernelBench-X: A Comprehensive Benchmark for Evaluating LLM-Generated GPU Kernels 7 upvotes, #23 of 2026-05-08
  25. The Granularity Axis: A Micro-to-Macro Latent Direction for Social Roles in Language Models 7 upvotes, #23 of 2026-05-08
  26. Prescriptive Scaling Laws for Data Constrained Training 6 upvotes, #26 of 2026-05-08
  27. PianoCoRe: Combined and Refined Piano MIDI Dataset 6 upvotes, #26 of 2026-05-08
  28. The Scaling Properties of Implicit Deductive Reasoning in Transformers 5 upvotes, #28 of 2026-05-08
  29. SwiftI2V: Efficient High-Resolution Image-to-Video Generation via Conditional Segment-wise Generation 5 upvotes, #28 of 2026-05-08
  30. When No Benchmark Exists: Validating Comparative LLM Safety Scoring Without Ground-Truth Labels 5 upvotes, #28 of 2026-05-08
  31. Recovering Hidden Reward in Diffusion-Based Policies 4 upvotes, #31 of 2026-05-08
  32. Generative Quantum-inspired Kolmogorov-Arnold Eigensolver 4 upvotes, #31 of 2026-05-08
  33. BioTool: A Comprehensive Tool-Calling Dataset for Enhancing Biomedical Capabilities of Large Language Models 4 upvotes, #31 of 2026-05-08
  34. Think, then Score: Decoupled Reasoning and Scoring for Video Reward Modeling 4 upvotes, #31 of 2026-05-08
  35. GeoStack: A Framework for Quasi-Abelian Knowledge Composition in VLMs 4 upvotes, #31 of 2026-05-08
  36. Are We Making Progress in Multimodal Domain Generalization? A Comprehensive Benchmark Study 4 upvotes, #31 of 2026-05-08
  37. EDU-CIRCUIT-HW: Evaluating Multimodal Large Language Models on Real-World University-Level STEM Student Handwritten Solutions 3 upvotes, #37 of 2026-05-08
  38. Sparkle: Realizing Lively Instruction-Guided Video Background Replacement via Decoupled Guidance 3 upvotes, #37 of 2026-05-08

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.