Skip to content
conftrace_
2025 ICML ICML 2025

MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking

Abstract

No abstract available. No abstract was captured for this paper when its proceedings were scraped - most often an older edition published without one online. How the corpus is built.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Reinforcement Learning

Authors

Sebastian Farquhar, Vikrant Varma, David Lindner, David Elson, Caleb Biddulph, Ian Goodfellow, Rohin Shah

Topics

Artificial Intelligence > Core AI > AI Safety Machine Learning > Optimization & Theory > Optimization Reinforcement Learning > Methods > Deep RL

Keywords

reinforcement learning large language model multi-step reward hacking myopic optimization non-myopic approval oversight
Download PDF