AutoRule: Reasoning Chain-of-thought Extracted Rule-based Rewards Improve Preference Learning figure
AlphaXiv 中文概览(可滚动查看)