图像是对齐的阿喀琉斯之踵：利用视觉漏洞越狱多模态大语言模型
Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models
AI总结：
该研究发现图像是多模态大语言模型的对齐漏洞，提出了利用精心制作的图像隐藏并放大文本恶意意图的HADES越狱方法，在主流模型上取得了较高的攻击成功率。
摘要：
本文研究多模态大语言模型（MLLMs）的无害性对齐问题。我们对代表性MLLMs的无害性表现开展了系统的实证分析，揭示出图像输入构成了MLLMs的对齐漏洞。受此启发，我们提出了一种名为HADES的新型越狱方法，该方法利用精心制作的图像，隐藏并放大文本输入中恶意意图的危害性。实验结果表明，HADES可有效越狱现有MLLMs，对LLaVA-1.5的平均攻击成功率（ASR）达90.26%，对Gemini Pro Vision达71.60%。我们的代码和数据可在https://github.com/RUCAIBox/HADES获取。
英文摘要：
In this paper, we study the harmlessness alignment problem of multimodal large language models (MLLMs). We conduct a systematic empirical analysis of the harmlessness performance of representative MLLMs and reveal that the image input poses the alignment vulnerability of MLLMs. Inspired by this, we propose a novel jailbreak method named HADES, which hides and amplifies the harmfulness of the malicious intent within the text input, using meticulously crafted images. Experimental results show that HADES can effectively jailbreak existing MLLMs, which achieves an average Attack Success Rate (ASR) of 90.26% for LLaVA-1.5 and 71.60% for Gemini Pro Vision. Our code and data are available at https://github.com/RUCAIBox/HADES.
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。
