Fetching the paper…

Honesty to Subterfuge: In-Context Reinforcement Learning Can Make Honest Models Reward Hack · Around