CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligence Paper • 2605.12882 • Published May 13 • 274
WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation Paper • 2605.10912 • Published May 11 • 46
Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction Paper • 2605.05242 • Published May 3 • 126
Personalized RewardBench: Evaluating Reward Models with Human Aligned Personalization Paper • 2604.07343 • Published Apr 8 • 13
Bootstrapping Exploration with Group-Level Natural Language Feedback in Reinforcement Learning Paper • 2603.04597 • Published Mar 4 • 211