Awesome RL with Human Feedback / 2024 / arXiv

Transforming and Combining Rewards for Aligning Large Language Models

作者：Zihao Wang, Chirag Nagpal, Jonathan Berant, Jacob Eisenstein, Alex D'Amour, Sanmi Koyejo, Victor Veitch · 单位：University of Chicago, Google Research, Google DeepMind, Stanford University · 来源：arXiv · 发布日期：2024

奖励规范奖励建模对齐目标设计
新标签打开 AlphaXiv
Transforming and Combining Rewards for Aligning Large Language Models figure
AlphaXiv 中文概览（可滚动查看）