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) and (ii) alignment via RL from human feedback (RLHF) (e.g., summarization in a human-preferred way). We give a new DP framework to achieve alignment via RL, and prove its correctness. Our experimental results validate the effectiveness of our approach, offering competitive utility while ensuring strong privacy protections.
@article{arxiv.2310.16960,
title = {Privately Aligning Language Models with Reinforcement Learning},
author = {Fan Wu and Huseyin A. Inan and Arturs Backurs and Varun Chandrasekaran and Janardhan Kulkarni and Robert Sim},
journal= {arXiv preprint arXiv:2310.16960},
year = {2024}
}