arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2401.12874cs.CLcs.AI
收藏

从理解到利用:大语言模型可解释性综述

From Understanding to Utilization: A Survey on Explainability for Large Language Models

Imperial College London(帝国理工学院)

Haoyan Luo, Lucia Specia

AI总结:

本文综述了大语言模型可解释性研究,从局部分析与全局分析两类方法出发,系统梳理了模型编辑、控制生成和模型增强等可解释性利用途径,并评估了相关指标与数据集,旨在连接理论理解与实际应用。

摘要:

大语言模型(LLMs)的可解释性是自然语言处理中一个关键但充满挑战的方面。随着LLMs日益成为各种应用不可或缺的组成部分,其“黑箱”特性引发了关于透明度和伦理使用的重大关切。本综述强调了提高LLMs可解释性的必要性,深入探讨了可解释性研究以及利用对这些模型的理解的各种方法和任务。我们的重点主要放在基于Transformer的预训练LLMs上,例如LLaMA系列,这些模型因其规模和复杂性而带来了独特的可解释性挑战。在现有方法方面,我们根据其解释目标将其分为局部分析和全局分析。在考虑可解释性的利用时,我们探索了几种引人注目的方法,这些方法集中于模型编辑、控制生成和模型增强。此外,我们还考察了具有代表性的评估指标和数据集,阐明了它们的优势和局限性。我们的目标是调和理论与实践理解与具体实现之间的关系,为LLMs时代的解释技术及其应用提出令人振奋的发展方向。

英文摘要:

Explainability for Large Language Models (LLMs) is a critical yet challenging aspect of natural language processing. As LLMs are increasingly integral to diverse applications, their "black-box" nature sparks significant concerns regarding transparency and ethical use. This survey underscores the imperative for increased explainability in LLMs, delving into both the research on explainability and the various methodologies and tasks that utilize an understanding of these models. Our focus is primarily on pre-trained Transformer-based LLMs, such as LLaMA family, which pose distinctive interpretability challenges due to their scale and complexity. In terms of existing methods, we classify them into local and global analyses, based on their explanatory objectives. When considering the utilization of explainability, we explore several compelling methods that concentrate on model editing, control generation, and model enhancement. Additionally, we examine representative evaluation metrics and datasets, elucidating their advantages and limitations. Our goal is to reconcile theoretical and empirical understanding with practical implementation, proposing exciting avenues for explanatory techniques and their applications in the LLMs era.

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。

↑