English

Curiosity-Driven Reinforcement Learning from Human Feedback

Computation and Language 2025-06-03 v2

Abstract

Reinforcement learning from human feedback (RLHF) has proven effective in aligning large language models (LLMs) with human preferences, but often at the cost of reduced output diversity. This trade-off between diversity and alignment quality remains a significant challenge. Drawing inspiration from curiosity-driven exploration in reinforcement learning, we introduce curiosity-driven RLHF (CD-RLHF), a framework that incorporates intrinsic rewards for novel states, alongside traditional sparse extrinsic rewards, to optimize both output diversity and alignment quality. We demonstrate the effectiveness of CD-RLHF through extensive experiments on a range of tasks, including text summarization and instruction following. Our approach achieves significant gains in diversity on multiple diversity-oriented metrics while maintaining alignment with human preferences comparable to standard RLHF. We make our code publicly available at https://github.com/ernie-research/CD-RLHF.

Keywords

Cite

@article{arxiv.2501.11463,
  title  = {Curiosity-Driven Reinforcement Learning from Human Feedback},
  author = {Haoran Sun and Yekun Chai and Shuohuan Wang and Yu Sun and Hua Wu and Haifeng Wang},
  journal= {arXiv preprint arXiv:2501.11463},
  year   = {2025}
}

Comments

ACL 2025

R2 v1 2026-06-28T21:11:18.678Z