Online Iterative Reinforcement Learning from Human Feedback with General Preference Model

Autor:	Ye, Chenlu, Xiong, Wei, Zhang, Yuheng, Dong, Hanze, Jiang, Nan, Zhang, Tong
Rok vydání:	2024
Předmět:	Computer Science - Machine Learning Statistics - Machine Learning
Druh dokumentu:	Working Paper
Popis:	We investigate Reinforcement Learning from Human Feedback (RLHF) in the context of a general preference oracle. In particular, we do not assume the existence of a reward function and an oracle preference signal drawn from the Bradley-Terry model as most of the prior works do. We consider a standard mathematical formulation, the reverse-KL regularized minimax game between two LLMs for RLHF under general preference oracle. The learning objective of this formulation is to find a policy so that it is consistently preferred by the KL-regularized preference oracle over any competing LLMs. We show that this framework is strictly more general than the reward-based one, and propose sample-efficient algorithms for both the offline learning from a pre-collected preference dataset and online learning where we can query the preference oracle along the way of training. Empirical studies verify the effectiveness of the proposed framework. Comment: RLHF, Preference Learning, Alignment for LLMs
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2402.07314 Zobrazit plný text záznamu View this record from Arxiv