作者位次 #2/8ACL 2026Poster· Natural Language Processing
HEAL ing Entropy Collapse: Enhancing Exploration in Few-Shot RLVR via Hybrid-Domain Entropy Dynamics Alignment
Zhanyu Liu · Qingguo Hu · Ante Wang · Chenqing Liu · Zhishang Xiang · Hui Li · Delai Qiu · Jinsong Su
摘要
Reinforcement Learning with Verifiable Reward (RLVR) has proven effective for training reasoning-oriented large language models, but existing methods largely assume high-resource settings with abundant training data. In low-resource scenarios, RLVR is prone to more severe entropy collapse, which...