Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignment

Published: 23 Oct 2023, Last Modified: 28 Nov 2023SoLaR PosterEveryoneRevisionsBibTeX
Keywords: Trustworthy LLM; Large Language Model; Responsible AI; Survey; Measurement Study; LLM Alignment
TL;DR: We present a comprehensive survey of key dimensions that are crucial to consider when assessing LLM trustworthiness along with measurement studies.
Abstract: Ensuring alignment has become a critical task before deploying large language models (LLMs) in real-world applications. A major challenge faced by practitioners is the lack of clear guidance on evaluating whether LLM outputs align with social norms, values, and regulations. This obstacle hinders the systematic iteration and deployment of LLMs. To address this issue, this paper presents a comprehensive survey of key dimensions that are crucial to consider when assessing LLM trustworthiness. The survey covers 7 major categories of LLM trustworthiness: reliability, safety, fairness, resistance to misuse, explainability and reasoning, adherence to social norms, and robustness. Each major category is further divided into several sub-categories, resulting in a total of 29 sub-categories. Additionally, a subset of 8 sub-categories is selected for further investigation, where corresponding measurement studies are designed and conducted on several widely-used LLMs. The measurement results indicate that, in general, more aligned models tend to perform better in terms of overall trustworthiness. However, the effectiveness of alignment varies across the different trustworthiness categories considered. This highlights the importance of conducting more fine-grained analyses, testing, and making continuous improvements on LLM alignment. By shedding light on these key dimensions of LLM trustworthiness, this paper aims to provide valuable insights and guidance to practitioners in the field. Understanding and addressing these concerns will be crucial in achieving reliable and ethically sound deployment of LLMs in various applications.
Submission Number: 86