AI.info

Rongyi Lin

Explore Rongyi Lin on AI.info.

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