作者位次 #9/10ACL 2026Findings· Findings of ACL
Rewarding the Rare: Uniqueness-Aware RL for Creative Problem Solving in LLM s
Zhiyuan Hu · Yucheng Wang · Yufei He · Jiaying Wu · Yilun Zhao · See-Kiong Ng · Cynthia Breazeal · Anh Tuan Luu · Hae Won Park · Bryan Hooi
摘要
Reinforcement learning (RL) has become a central paradigm for post-training large language models (LLMs), particularly for complex reasoning tasks, yet it often suffers from exploration collapse: policies prematurely concentrate on a small set of dominant reasoning patterns, improving pass@1 whil...