2310.12931

N/A

Published as a conference paper at ICLR 2024 E UREKA : H UMAN - LEVEL R EWARD D ESIGN VIA C ODING L ARGE L ANGUAGE M ODELS Yecheng Jason Ma1 2 , William Liang2 , Guanzhi Wang1 3 , De-An Huang1 , Osbert Bastani2 , Dinesh Jayaraman2 , Yuke Zhu1 4 , Linxi “Jim” Fan1 † , Anima Anandkumar1 3 † 1 NVIDIA, 2 UPenn, 3…

related reading

• 2310.12921.pdfarxiv.org
• [2305.18290] Direct Preference Optimization: Your Language Model is Secretly a Reward Modelarxiv.org
• [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org
• 2401.10020.pdfarxiv.org
• [2507.19457] GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningarxiv.org
• [2412.14135] Scaling of Search and Learning: A Roadmap to Reproduce o1 from Reinforcement Learning Perspectivearxiv.org
• 2307.12950.pdfarxiv.org
• Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domainsarxiv.org
• A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulationtoyotaresearchinstitute.github.io
• 2310.01405.pdfarxiv.org
• 2310.01405arxiv.org
• [2603.21972] Demystifying Reinforcement Learning for Long-Horizon Tool-Using Agents: A Comprehensive Recipearxiv.org