Large Language Models for Explainability in Machine Learning
Keywords: XAI, explainability, large language models
TL;DR: A quantitative evaluation of the suitability of LLMs for explainable AI.
Abstract: We investigate the potential of large language models (LLMs) in explainable artificial intelligence (XAI) by examining their ability to generate understandable explanations for machine learning (ML) models. While recent studies suggest that LLMs could effectively address the limitations of traditional explanation methods through their conversational capabilities, there has been a lack of systematic evaluation of the quality of these LLM-generated explanations. To fill this gap, this study evaluates whether LLMs can produce explanations for ML models that meet the fundamental properties of XAI using conventional ML models and explanation methods as benchmarks. The findings offer important insights into the strengths and limitations of LLMs as tools for explainable AI, provide recommendations for their appropriate use, and identify promising directions for future research.
Primary Area: interpretability and explainable AI
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 10132
Loading