Jailbreak Connectivity：面向多模态大语言模型的多样化、可迁移、通用越狱框架
Jailbreak Connectivity: Towards Diverse, Transferable, and Universal Multimodal Large Language Model Jailbreak
AI 导读
研究者提出 Jailbreak Connectivity（JC）框架，通过在图像空间低损失区域优化连续对抗路径，实现多样化、可迁移且通用的多模态大语言模型越狱攻击。
正文 · 原文
Abstract:Multimodal large language models (MLLMs) are vulnerable to visual jailbreak attacks that bypass safety safeguards. Existing methods typically optimize isolated adversarial examples, limiting attack diversity, transferability, and generalization. We propose Jailbreak Connectivity (JC), a framework based on the observation that effective adversarial inputs can be connected through low-loss regions in image space. JC optimizes continuous adversarial paths: path sampling yields diverse jailbreak images, lightweight surrogate classifiers guide cross-model transfer, and a distribution-level objective enables universal attacks across harmful queries. On MM-SafetyBench, JC achieves average attack success rates of 84.01%, 85.77%, and 97.95% on Qwen3-VL-8B-Instruct, InternVL3.5-8B, and Gemma-3-12B-IT, respectively, exceeding the strongest baselines by up to 68.36 percentage points. In black-box evaluations, JC improves attack success rates over FigStep by 24.69 and 27.34 percentage points on GPT-5.6 and Gemini 3.1 Pro, respectively. These findings reveal structured vulnerabilities in MLLM safety alignment and inform more comprehensive security evaluation.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR) |
| Cite as: | arXiv:2610.38091 [cs.LG] |
| (or arXiv:2610.38091v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.38091 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Xu Zhang [view email]
[v1]
Wed, 28 Oct 2026 00:13:19 UTC (6,662 KB)
来源：arXiv cs.AI · arxiv.org