Privately Aligning Language Models with Reinforcement Learning

arXiv:2310.16960v2 Announce Type: replace
Abstract: Positioned between pre-training and user deployment, aligning large language models (LLMs) through reinforcement learning (RL) has emerged as a prevailing strategy for training instruction following-models such as ChatGPT. In this work, we initiate the study of privacy-preserving alignment of LLMs through Differential Privacy (DP) in conjunction with RL. Following the influential work of Ziegler et al. (2020), we study two dominant paradigms: (i) alignment via RL without human in the loop (e.g., positive review generation) …

abstract alignment arxiv chatgpt cs.cr cs.lg deployment differential differential privacy language language models large language large language models llms pre-training privacy reinforcement reinforcement learning strategy study through training type work