AI.info

Junwei He

Explore Junwei He on AI.info.

• Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics