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 2407.09292cs.CR
收藏

大语言模型上的反事实可解释增量提示攻击分析

Counterfactual Explainable Incremental Prompt Attack Analysis on Large Language Models

Dong Shu, Mingyu Jin, Tianle Chen, Chong Zhang, Yongfeng Zhang

AI总结:

针对大语言模型的安全漏洞,提出反事实可解释增量提示攻击方法CEIPA,通过四级增量修改定量评估攻击效果并揭示有害响应成因,显著提升攻击提示有效性。

摘要:

本研究阐明了通过可解释的提示攻击分析来识别和缓解大型语言模型(LLMs)如GPT-4和LLaMA-2中的漏洞,从而加强其安全性和隐私保护措施的迫切需求。我们提出了反事实可解释增量提示攻击(CEIPA),这是一种新技术,通过特定方式引导提示来定量衡量攻击有效性并探索这些模型中内置的防御机制。我们的方法独特之处在于能够通过增量反事实方法阐明LLMs生成有害响应的原因。通过将提示修改过程组织为四个增量级别:(词、句、字符以及字符与词的组合),我们促进了对LLMs固有脆弱性的全面检查。我们的研究结果不仅提供了反事实解释的洞察,还证明了我们的框架显著增强了攻击提示的有效性。

英文摘要:

This study sheds light on the imperative need to bolster safety and privacy measures in large language models (LLMs), such as GPT-4 and LLaMA-2, by identifying and mitigating their vulnerabilities through explainable analysis of prompt attacks. We propose Counterfactual Explainable Incremental Prompt Attack (CEIPA), a novel technique where we guide prompts in a specific manner to quantitatively measure attack effectiveness and explore the embedded defense mechanisms in these models. Our approach is distinctive for its capacity to elucidate the reasons behind the generation of harmful responses by LLMs through an incremental counterfactual methodology. By organizing the prompt modification process into four incremental levels: (word, sentence, character, and a combination of character and word) we facilitate a thorough examination of the susceptibilities inherent to LLMs. The findings from our study not only provide counterfactual explanation insight but also demonstrate that our framework significantly enhances the effectiveness of attack prompts.

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。

↑