2025
MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking
ICML 2025poster
Future advanced AI systems may learn sophisticated strategies through reinforcement learning (RL) that humans cannot understand well enough to safely evaluate. We propose a training method which avoids agents learning undesired multi-step plans that receive high reward (multi-step "reward hacks") ev…