作者位次 #4/6ACL 2026Findings· Findings of ACL
NSF - C o T : Neuro-Symbolic Formal Verification of Chain-of-Thought Faithfulness in Contextual Question Answering
Vishal Pramanik · Maisha Maliha · Nathaniel D. Bastian · Alvaro Velasquez · Susmit Jha · Sumit Kumar Jha
摘要
Chain-of-thought (CoT) prompting makes language models write step-by-step explanations, but these steps may not match what the model actually used to choose its answer. Existing faithfulness checks often only test whether changing the written chain changes the answer, without verifying whether th...