Awesome RL with Human Feedback / 2024 / arXiv

Reward-Augmented Data Enhances Direct Preference Alignment of LLMs

作者：Shenao Zhang, Zhihan Liu, Boyi Liu, Yufeng Zhang, Yingxiang Yang, Yongfei Liu, Liyu Chen, Tao Sun, Zhaoran Wang · 来源：arXiv · 发布日期：2024

奖励建模偏好学习对齐目标设计策略优化策略梯度人类反馈强化学习
新标签打开 AlphaXiv新标签打开 Code
Reward-Augmented Data Enhances Direct Preference Alignment of LLMs figure
AlphaXiv 中文概览（可滚动查看）