Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics
Paper • 2609.38847 • Published • 11
daily-paper
-
-
Agent-Reflex: Teaching Language Agents to Act Reflectively
Paper • 2610.17953 • Published • 36 -
Multiscale Policy Optimization for Long-Horizon LLM Agents
Paper • 2610.35013 • Published • 22 -
The Trinity:Capability-Driven Recursive Self-Improvement for Multimodal Agents
Paper • 2610.36077 • Published • 165 -
Memory Tax: Dissecting the Net Value of Agent Memory
Paper • 2610.34073 • Published • 25