BAITBENCH: Measuring Agent Reward Hacking with Optional Shortcuts Planted in ML Tasks
This work releases BAITBENCH, a suite of three synthetic tabular ML tasks that each contain a shortcut that allows agents to inflate the public test score but fail on a hidden test set, and releases an annotated dataset of transcripts containing reward hacks as a testbed for evaluating reward-hacking mitigations head-to-head.
Pradyumna Shyama Prasad, M. Anto, Leon Eshuijs et al.
· 1 citation