Research
Rubric-Based RL Reward Hacking: What Two New Papers Claim
Two new preprints on rubric-based RL reward hacking study how models game LLM judges, and one title says higher scores can come with worse answers.