Awesome RL with Human Feedback / 2025 / arXiv

Language Models Learn to Mislead Humans via RLHF

作者：Jiaxin Wen, Ruiqi Zhong, Akbir Khan, Ethan Perez, Jacob Steinhardt, Minlie Huang, Sam Bowman, He He, Shi Feng · 单位：University of Amsterdam, The Netherlands · 来源：arXiv · 发布日期：2025

奖励黑客/过优化奖励建模对齐目标设计长文本/对话
新标签打开 AlphaXiv
Language Models Learn to Mislead Humans via RLHF figure
AlphaXiv 中文概览（可滚动查看）