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 2511.08367cs.CR
收藏

为什么弱分布外(Weak-OOD)样本有帮助?迈向理解视觉语言模型越狱的更深入一步

Why Does Weak-OOD Help? A Further Step Towards Understanding Jailbreaking VLMs

Tsinghua Shenzhen International Graduate School, Tsinghua University(清华大学深圳国际研究生院);ByteDance(字节跳动);Peking University(北京大学);Nanyang Technological University(南洋理工大学);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);College of Computer Science and Software Engineering, Shenzhen University(深圳大学计算机科学与软件工程学院)

Yuxuan Zhou, Yuzhao Peng, Yang Bai, Kuofeng Gao, Yihao Zhang, Yechao Zhang, Xun Chen, Tao Yu, Tao Dai, Shu-Tao Xia

AI总结:

该研究揭示了视觉语言模型越狱中的弱-OOD非单调现象,通过输入意图感知与模型拒绝触发的权衡解释该现象,并受OCR能力启发设计JOCR方法,实现了更优的平均越狱攻击成功率。

摘要:

大型视觉语言模型(VLM)易受越狱攻击:研究人员已开发出多种绕过VLM安全机制的攻击策略。在这些方法中,基于分布外(OOD)策略的越狱方法因简单有效而受到广泛关注。本文进一步深化了对基于OOD的VLM越狱方法的理解。我们发现,轻度OOD操纵能实现比干净输入和过强扰动更强的越狱性能,我们将这种非单调模式定义为“弱-OOD(weak-OOD)”。我们通过两个主导因素之间的权衡来解释这一现象:输入意图感知与模型拒绝触发。我们的证据表明,这两个因素对OOD操纵的响应不同,这与广泛预训练鲁棒性和较窄安全对齐之间的差异相一致。基于这一见解,我们从主流VLM预训练阶段的核心任务——光学字符识别(OCR)能力增强中汲取灵感。利用这一能力,我们设计了JOCR(基于OCR感知的嵌入文本扰动越狱),这是嵌入文本越狱的一种实用OCR可读扩展方案,在评估的基线方法中取得了最优的平均攻击成功率(ASR)。代码可在GitHub获取:https://github.com/Yuxuan2003/weak-ood-jailbreak。

英文摘要:

Large Vision-Language Models (VLMs) are susceptible to jailbreak attacks: researchers have developed various attack strategies that bypass the safety mechanisms of VLMs. Among these approaches, jailbreak methods based on the Out-of-Distribution (OOD) strategy have garnered widespread attention due to their simplicity and effectiveness. This paper further advances the understanding of OOD-based VLM jailbreak methods. We show that mild OOD manipulations can achieve stronger jailbreak performance than both clean inputs and overly strong perturbations, a non-monotonic pattern we define as "weak-OOD". We explain this phenomenon through a trade-off between two dominant factors: input intent perception and model refusal triggering. Our evidence suggests that these two factors respond differently to OOD manipulations, which is consistent with a discrepancy between broad pre-training robustness and narrower safety alignment. Building on this insight, we draw inspiration from optical character recognition (OCR) capability enhancement---a core task in the pre-training phase of mainstream VLMs. Leveraging this capability, we design JOCR (Jailbreak via OCR-Aware Embedded Text Perturbation), a practical OCR-readable extension of embedded-text jailbreaks that achieves the best average ASR among evaluated baselines. Code is available at GitHub: https://github.com/Yuxuan2003/weak-ood-jailbreak.

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。

↑