AUG 12, 2026 · PREPRINT
Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL
arXiv
This is a preprint describing an unblinded, single-model empirical study of a proposed mitigation method for reward hacking, lacking peer review, independent replication, or comparison to published baselines.
Reported
Gold judge score drop (HealthBenc…3 points
Gold judge score drop (ResearchQA)22 points
OOD gold score improvement at 30%…+1 to +2 points