[2603.21208] JANUS: A Lightweight Framework for Jailbreaking Text-to-Image Models via Distribution Optimization

Computer Science > Computer Vision and Pattern Recognition arXiv:2603.21208 (cs) [Submitted on 22 Mar 2026] Title:JANUS: A Lightweight Framework for Jailbreaking Text-to-Image Models via Distribution Optimization Authors:Haolun Zheng, Yu He, Tailun Chen, Shuo Shao, Zhixuan Chu, Hongbin Zhou, Lan Tao, Zhan Qin, Kui Ren View a PDF of the paper titled JANUS: A Lightweight Framework for Jailbreaking Text-to-Image Models via Distribution Optimization, by Haolun Zheng and 8 other authors View PDF HTML (experimental) Abstract:Text-to-image (T2I) models such as Stable Diffusion and DALLE remain susceptible to generating harmful or Not-Safe-For-Work (NSFW) content under jailbreak attacks despite deployed safety filters. Existing jailbreak attacks either rely on proxy-loss optimization instead of the true end-to-end objective, or depend on large-scale and costly RL-trained generators. Motivated by these limitations, we propose JANUS , a lightweight framework that formulates jailbreak as optimizing a structured prompt distribution under a black-box, end-to-end reward from the T2I system and its safety filters. JANUS replaces a high-capacity generator with a low-dimensional mixing policy over two semantically anchored prompt distributions, enabling efficient exploration while preserving the target semantics. On modern T2I models, we outperform state-of-the-art jailbreak methods, improving ASR-8 from 25.30% to 43.15% on Stable Diffusion 3.5 Large Turbo with consistently higher CLIP and...