Tag
This paper introduces an LLM-as-a-judge method to measure perturbation strength for assessing self-consistency in LLM explanations, showing that input perturbations generally affect LLMs more strongly than CoT perturbations.