PERL: Parameter Efficient Reinforcement Learning from Human Feedback

Hakim Sidahmed; Samrat Phatale; Alex Hutcheson; Zhuonan Lin; Zhang Chen; Zac Yu; Jarvis Jin; Roman Komarytsia; Christiane Ahlheim; Yonghao Zhu; Simral Chaudhary; Bowen Li; Saravanan Ganesh; Bill Byrne; Jessica Hoffmann; Hassan Mansoor; Wei Li; Abhinav Rastogi; Lucas Dixon

Paper · arXiv 2403.10704

PERL: Parameter Efficient Reinforcement Learning from Human Feedback

Hakim Sidahmed, Samrat Phatale, Alex Hutcheson, Zhuonan Lin, Zhang Chen, Zac Yu, Jarvis Jin, Roman Komarytsia, Christiane Ahlheim, Yonghao Zhu, Simral Chaudhary, Bowen Li, Saravanan Ganesh, Bill Byrne, Jessica Hoffmann, Hassan Mansoor, Wei Li, Abhinav Rastogi, Lucas Dixon

60 upvotesMarch 15, 2024arXiv 预印本

AI 摘要

Parameter-efficient reinforcement learning using LoRA in RLHF matches conventional RLHF performance with reduced computational cost and memory usage.

Reinforcement Learning from Human FeedbackPretrained Large Language ModelsLow-Rank AdaptationParameter Efficient Reinforcement Learningreward model trainingreinforcement learningfull-tuningTaskmaster CoffeeTaskmaster Ticketing

Abstract

Reinforcement Learning from Human Feedback (RLHF) has proven to be a strong method to align Pretrained Large Language Models (LLMs) with human preferences. But training models with RLHF is computationally expensive, and an overall complex process. In this work, we study RLHF where the underlying models are trained using the parameter efficient method of Low-Rank Adaptation (LoRA) introduced by Hu et al. [2021]. We investigate the setup of "Parameter Efficient Reinforcement Learning" (PERL), in which we perform reward model training and reinforcement learning using LoRA. We compare PERL to conventional fine-tuning (full-tuning) across various configurations for 7 benchmarks, including 2 novel datasets, of reward modeling and reinforcement learning. We find that PERL performs on par with the conventional RLHF setting, while training faster, and with less memory. This enables the high performance of RLHF, while reducing the computational burden that limits its adoption as an alignment technique for Large Language Models. We also release 2 novel thumbs up/down preference datasets: "Taskmaster Coffee", and "Taskmaster Ticketing" to promote research around RLHF.