تقرير
Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models
العنوان: | Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models |
---|---|
المؤلفون: | Li, Yifan, Guo, Hangyu, Zhou, Kun, Zhao, Wayne Xin, Wen, Ji-Rong |
سنة النشر: | 2024 |
المجموعة: | Computer Science |
مصطلحات موضوعية: | Computer Science - Computer Vision and Pattern Recognition, Computer Science - Computation and Language |
الوصف: | 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 will be publicly released. Comment: Work in progress |
نوع الوثيقة: | Working Paper |
الوصول الحر: | http://arxiv.org/abs/2403.09792Test |
رقم الانضمام: | edsarx.2403.09792 |
قاعدة البيانات: | arXiv |
الوصف غير متاح. |