Question: Reward Hacking Detection Metrics in RL-Trained Agents

Anomaly detection vs. reward auditing for hacking