AIJanuary 26, 2025Google DeepMind Introduces MONA: A Novel Machine Learning Framework to Mitigate Multi-Step Reward Hacking in Reinforcement Learning
