Agent-ValueBench: A Comprehensive Benchmark for Evaluating Agent Values Paper • 2605.10365 • Published 6 days ago • 9
KL for a KL: On-Policy Distillation with Control Variate Baseline Paper • 2605.07865 • Published 9 days ago • 17
Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States Paper • 2605.07579 • Published 9 days ago • 15