Chain-of-thought (CoT) prompting has been shown to empirically improve the
accuracy of large language models (LLMs) on various question answering tasks.
While understanding why CoT prompting is effective is crucial to ensuring that
this phenomenon is a consequence of desired model behavior, little work has
addressed this; nonetheless, such an understanding is a critical prerequisite
for responsible model deployment. We address this question by leveraging
gradient-based feature attribution methods which produce saliency scores that
capture the influence of input tokens on model output. Specifically, we probe
several open-source LLMs to investigate whether CoT prompting affects the
relative importances they assign to particular input tokens. Our results
indicate that while CoT prompting does not increase the magnitude of saliency
scores attributed to semantically relevant tokens in the prompt compared to
standard few-shot prompting, it increases the robustness of saliency scores to
question perturbations and variations in model output.

CoT 提示法通过梯度特征归因方法揭示其对模型输入的影响，提高了模型对问题扰动和输出变化的鲁棒性。