When the Chain Breaks: Interactive Diagnosis of LLM Chain-of-Thought Reasoning Errors

Shiwei Chen, Niruthikka Sritharan, Xiaolin Wen +3

Current Large Language Models (LLMs), especially Large Reasoning Models, can generate Chain-of-Thought (CoT) reasoning traces to illustrate how they produce final outputs, thereby…