大语言模型的自解释是否具有忠实性？
Are self-explanations from Large Language Models faithful?
AI总结：
针对大语言模型自解释可能误导信任的问题，提出用自一致性检验衡量其可解释性忠实度，首次将该方法应用于三类自解释场景，发现忠实度随解释类型、模型和任务变化，自解释总体不可轻信。
摘要：
经过指令微调的大语言模型（Large Language Models, LLMs）在众多任务上表现出色，甚至能解释自身的推理过程，即所谓的自解释。然而，具有说服力但错误的自解释会让人们对LLMs产生无依据的信任，进而增加风险。因此，衡量自解释是否真实反映模型行为至关重要，这类衡量标准被称为可解释性忠实度，由于无法获取真实依据，且许多LLM仅提供推理API，该衡量工作颇具挑战性。为解决这一问题，我们提出采用自一致性检验来衡量忠实度。例如，如果LLM称某一组词对做出预测很重要，那么去掉这些词后它就无法做出该预测。尽管自一致性检验是衡量忠实度的常用方法，但此前尚未成功应用于LLM在反事实、特征归因和删改解释三类场景下的自解释。我们的结果表明，忠实度取决于解释类型、模型和任务，说明自解释总体上不可信。例如在情感分类任务中，反事实解释对Llama2更忠实，特征归因对Mistral更忠实，删改解释对Falcon 40B更忠实。
英文摘要：
Instruction-tuned Large Language Models (LLMs) excel at many tasks and will even explain their reasoning, so-called self-explanations. However, convincing and wrong self-explanations can lead to unsupported confidence in LLMs, thus increasing risk. Therefore, it's important to measure if self-explanations truly reflect the model's behavior. Such a measure is called interpretability-faithfulness and is challenging to perform since the ground truth is inaccessible, and many LLMs only have an inference API. To address this, we propose employing self-consistency checks to measure faithfulness. For example, if an LLM says a set of words is important for making a prediction, then it should not be able to make its prediction without these words. While self-consistency checks are a common approach to faithfulness, they have not previously been successfully applied to LLM self-explanations for counterfactual, feature attribution, and redaction explanations. Our results demonstrate that faithfulness is explanation, model, and task-dependent, showing self-explanations should not be trusted in general. For example, with sentiment classification, counterfactuals are more faithful for Llama2, feature attribution for Mistral, and redaction for Falcon 40B.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
