Daily Papers of 2025-10-24

  1. Human-Agent Collaborative Paper-to-Page Crafting for Under $0.1 66 upvotes, #1 of 2025-10-24
  2. AdaSPEC: Selective Knowledge Distillation for Efficient Speculative Decoders 58 upvotes, #2 of 2025-10-24
  3. Open-o3 Video: Grounded Video Reasoning with Explicit Spatio-Temporal Evidence 52 upvotes, #3 of 2025-10-24
  4. HoloCine: Holistic Generation of Cinematic Multi-Shot Long Video Narratives 38 upvotes, #4 of 2025-10-24
  5. DyPE: Dynamic Position Extrapolation for Ultra High Resolution Diffusion 31 upvotes, #5 of 2025-10-24
  6. Loopholing Discrete Diffusion: Deterministic Bypass of the Sampling Wall 22 upvotes, #6 of 2025-10-24
  7. SAKE: Towards Editing Auditory Attribute Knowledge of Large Audio-Language Models 19 upvotes, #7 of 2025-10-24
  8. Every Question Has Its Own Value: Reinforcement Learning with Explicit Human Values 18 upvotes, #8 of 2025-10-24
  9. Investigating Safety Vulnerabilities of Large Audio-Language Models Under Speaker Emotional Variations 17 upvotes, #9 of 2025-10-24
  10. The Massive Legal Embedding Benchmark (MLEB) 17 upvotes, #9 of 2025-10-24
  11. Seed3D 1.0: From Images to High-Fidelity Simulation-Ready 3D Assets 17 upvotes, #9 of 2025-10-24
  12. Search Self-play: Pushing the Frontier of Agent Capability without Supervision 15 upvotes, #12 of 2025-10-24
  13. Thought Communication in Multiagent Collaboration 13 upvotes, #13 of 2025-10-24
  14. Conan: Progressive Learning to Reason Like a Detective over Multi-Scale Visual Evidence 11 upvotes, #14 of 2025-10-24
  15. Diff-XYZ: A Benchmark for Evaluating Diff Understanding 8 upvotes, #15 of 2025-10-24
  16. ARGenSeg: Image Segmentation with Autoregressive Image Generation Model 8 upvotes, #15 of 2025-10-24
  17. LayerComposer: Interactive Personalized T2I via Spatially-Aware Layered Canvas 8 upvotes, #15 of 2025-10-24
  18. CiteGuard: Faithful Citation Attribution for LLMs via Retrieval-Augmented Validation 7 upvotes, #18 of 2025-10-24
  19. AlphaFlow: Understanding and Improving MeanFlow Models 7 upvotes, #18 of 2025-10-24
  20. Scaling Laws Meet Model Architecture: Toward Inference-Efficient LLMs 6 upvotes, #20 of 2025-10-24
  21. From Masks to Worlds: A Hitchhiker's Guide to World Models 6 upvotes, #20 of 2025-10-24
  22. ImpossibleBench: Measuring LLMs' Propensity of Exploiting Test Cases 5 upvotes, #22 of 2025-10-24
  23. Long-Context Attention Benchmark: From Kernel Efficiency to Distributed Context Parallelism 4 upvotes, #23 of 2025-10-24
  24. Adamas: Hadamard Sparse Attention for Efficient Long-Context Inference 4 upvotes, #23 of 2025-10-24
  25. MSC-Bench: A Rigorous Benchmark for Multi-Server Tool Orchestration 4 upvotes, #23 of 2025-10-24
  26. Communication to Completion: Modeling Collaborative Workflows with Intelligent Multi-Agent Communication 4 upvotes, #23 of 2025-10-24
  27. Emergence of Linear Truth Encodings in Language Models 2 upvotes, #27 of 2025-10-24
  28. ComProScanner: A multi-agent based framework for composition-property structured data extraction from scientific literature 2 upvotes, #27 of 2025-10-24

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.