Back to papers
June 8, 2026cs.AIcs.LG

Proxy Reward Internalization and Mechanistic Exploitation: A Learned Precursor to Reward Hacking and Its Generalization

Categories

cs.AI, cs.LG