Daily Papers of 2026-09-14

  1. Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation 170 upvotes, #1 of 2026-09-14
  2. DataFlex-RL: An Evaluation Platform for RLVR Data Policies 162 upvotes, #2 of 2026-09-14
  3. Feyospace-v1: How the Cyber Mercury Seven Trained Frontier Cyber Models 124 upvotes, #3 of 2026-09-14
  4. StepAudio 3 Gen Technical Report 99 upvotes, #4 of 2026-09-14
  5. Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-work 93 upvotes, #5 of 2026-09-14
  6. Breaking the Vision-Action Shortcut: Latent Interface Training for Generalizable Robotics Foundation Models 70 upvotes, #6 of 2026-09-14
  7. SAS: Simple Attention Sparsification via End-to-End Optimization of Context Ranking 65 upvotes, #7 of 2026-09-14
  8. ReactHuman: A Physics-Grounded Benchmark for Human-Like Reactive Decision-Making in Embodied Multimodal LLMs 52 upvotes, #8 of 2026-09-14
  9. COBRA-Skills: Contextual Bandit-Guided Evolution for Agent Skill Optimization 43 upvotes, #9 of 2026-09-14
  10. PLC-DPO: Posterior Label Correction in Noisy and Ambiguous Preference Optimization 27 upvotes, #10 of 2026-09-14
  11. Pelican-Sim 1.0: A General World Model Simulator for Embodied Intelligence 22 upvotes, #11 of 2026-09-14
  12. ActionSplice: In-Flight Action Editing for Interactive World Models 10 upvotes, #12 of 2026-09-14
  13. MobileVLA-R1 2.0: RL-Enhanced Reasoning for Mobile Robot Control 8 upvotes, #13 of 2026-09-14
  14. Beyond Top-k Skill Retrieval: Diversity-Aware Skill Routing for LLM Agents 7 upvotes, #14 of 2026-09-14
  15. Online Learning with LLM Experts from Limited Feedback 6 upvotes, #15 of 2026-09-14
  16. Studying Without a Syllabus: Task-Agnostic Environment Preprocessing 4 upvotes, #16 of 2026-09-14
  17. How Far Can Synthetic Data Take Thai OCR? 3 upvotes, #17 of 2026-09-14
  18. Ambient @ EgoProactive 2026 : Proactive Egocentric Assistance with Visually Grounded Supervision 3 upvotes, #17 of 2026-09-14
  19. Ambient @ EgoLongQA 2026: Distilling Long-Video perception into a Sub-2B Model 3 upvotes, #17 of 2026-09-14
  20. TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents 3 upvotes, #17 of 2026-09-14
  21. SNAP3D: Physically Grounded 3D Parts for Assembly from a Single Image 3 upvotes, #17 of 2026-09-14
  22. Towards a Deterministic Math Solver for Clinical Language Models 2 upvotes, #22 of 2026-09-14
  23. Competence-Gated Pooling of Language Models and Priors for Event Forecasting 2 upvotes, #22 of 2026-09-14
  24. Building and Evaluating Fixed-Voice Thai TTS from Synthetic Speech 1 upvotes, #24 of 2026-09-14
  25. ReMoMask-2: Latent Retrieval-Augmented Masked Motion Generation 1 upvotes, #24 of 2026-09-14
  26. Feature Recovery for Object Understanding After Irreversible Fire Damage 1 upvotes, #24 of 2026-09-14

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.