LifeSearch: Benchmarking Deep Research Agents for Real-World Life and Productivity Paper • 2610.18092 • Published 8 days ago • 21
LongR: Unleashing Long-Context Reasoning via Reinforcement Learning with Dense Utility Rewards Paper • 2602.05758 • Published 9 days ago • 27
AReA: Adaptive Receptive-Field Attention for Efficient High-Resolution Diffusion Transformers Paper • 2610.17227 • Published 9 days ago • 58
PhyProbe: Rethinking Physical Consistency Evaluation in Generated Videos Paper • 2609.38377 • Published 22 days ago • 17
FlashVSR+: Efficient Video Super-Resolution via Hybrid Attention and Reward-Yield Distribution Matching Paper • 2610.17552 • Published 8 days ago • 45
Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics Paper • 2609.38847 • Published 21 days ago • 69