图像到文本逻辑越狱：你的想象力能帮你做任何事
Image-to-Text Logic Jailbreak: Your Imagination can Help You Do Anything
AI总结：
本文针对大型视觉语言模型在流程图逻辑理解方面的越狱漏洞，提出了新数据集Flow-JD评估图像到文本逻辑越狱能力，发现GPT-4o等模型越狱率高达92.8%，揭示了VLMs的安全隐患并呼吁开发有效防御。
摘要：
诸如GPT-4V等大型视觉语言模型在生成全面且细致的回复方面取得了显著成功。研究人员提出了各种基准来评估VLMs的能力。随着VLMs中视觉和文本输入的整合，新的安全问题随之出现，因为恶意攻击者可以利用多种模态来实现其目标。这导致人们越来越关注VLMs在越狱方面的漏洞。现有的大多数研究集中在生成对抗性图像或无意义图像来越狱这些模型。然而，没有研究人员评估VLMs在流程图中的逻辑理解能力是否会影响越狱。因此，为了填补这一空白，本文首先引入了一个新数据集Flow-JD，专门设计用于评估VLMs基于逻辑的流程图越狱能力。我们在GPT-4o、GPT-4V及其他5个SOTA开源VLMs上进行了广泛评估，越狱率高达92.8%。我们的研究揭示了当前VLMs在图像到文本越狱方面的重大漏洞，这些发现凸显了未来开发稳健有效防御措施的紧迫性。
英文摘要：
Large Visual Language Model\textbfs (VLMs) such as GPT-4V have achieved remarkable success in generating comprehensive and nuanced responses. Researchers have proposed various benchmarks for evaluating the capabilities of VLMs. With the integration of visual and text inputs in VLMs, new security issues emerge, as malicious attackers can exploit multiple modalities to achieve their objectives. This has led to increasing attention on the vulnerabilities of VLMs to jailbreak. Most existing research focuses on generating adversarial images or nonsensical image to jailbreak these models. However, no researchers evaluate whether logic understanding capabilities of VLMs in flowchart can influence jailbreak. Therefore, to fill this gap, this paper first introduces a novel dataset Flow-JD specifically designed to evaluate the logic-based flowchart jailbreak capabilities of VLMs. We conduct an extensive evaluation on GPT-4o, GPT-4V, other 5 SOTA open source VLMs and the jailbreak rate is up to 92.8%. Our research reveals significant vulnerabilities in current VLMs concerning image-to-text jailbreak and these findings underscore the the urgency for the development of robust and effective future defenses.
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。
