作者位次 #3/6ACL 2026Findings· Findings of ACL
Tracing the Light of Thought: A Probabilistic Self- and Cross-Consistency Verification Mechanism Improving Mathematical Reasoning in LLM s
Xiaoyang Liu · Dawei Wang · Tian Li · Huizhi Liang · Gary Ushaw · Richard Davison
摘要
Reasoning capability is fundamental in enabling Large Language Models to perform complex multi-step inference. By sampling multiple reasoning paths and selecting the most frequent answer, Self Consistency (SC) remains highly effective but fails on challenging tasks where incorrect answers dominat...