Abstract

Large language models (LLMs) have revolutionized artificial intelligence, but their increasing deployment across critical domains has raised concerns about their abnormal behaviors when faced with malicious attacks. Such vulnerability alerts the widespread inadequacy of pre-release testing. In this paper, we conduct a comprehensive empirical study to evaluate the effectiveness of traditional coverage criteria in identifying such inadequacies, exemplified by the significant security concern of jailbreak attacks. Our study begins with a clustering analysis of the hidden states of LLMs, revealing that the embedded characteristics effectively distinguish between different query types. We then systematically evaluate the performance of these criteria across three key dimensions: criterion level, layer level, and token level.

Our research uncovers significant differences in the sets of neurons covered when LLMs process normal versus jailbreak queries, aligning with our clustering experiments. Leveraging these findings, we propose three practical applications of coverage criteria in the context of LLM security testing. Specifically, we develop a real-time jailbreak detection mechanism that achieves high accuracy (93.61% on average) in classifying queries as normal or jailbreak. Furthermore, we explore the use of coverage levels to prioritize test cases, improving testing efficiency by focusing on high-risk interactions and removing redundant tests. Lastly, we introduce a coverage-guided approach for generating jailbreak attack examples, enabling systematic refinement of prompts to uncover new vulnerabilities. This study improves our understanding of LLM security testing, enhances their safety, and provides a foundation for developing more robust AI applications.

Overview

Source Code Demo

https://anonymous.4open.science/r/Empirical-Study-of-Coverage-Criteria-7EDC