作者位次 #3/4ACL 2026Poster· Natural Language Processing
Optimizing Length Compression in Large Reasoning Models
Zhengxiang Cheng · Dongping Chen · Mingyang Fu · Tianyi Zhou
摘要
Large Reasoning Models (LRMs) have achieved remarkable success, yet they often suffer from producing unnecessary and verbose reasoning chains. We identify a core aspect of this issue as ”invalid thinking”— models tend to repeatedly double-check their work after having derived the correct answer....