This work proposes AEGIS (Adaptive Evasion Guard via Identification and Steering), an inference-time defense that applies similarity-aware repulsion only at the identified vulnerable heads of VSA and preserves benign fidelity, avoids suppressing hard-negative concepts, and transfers to SD 2.1 and FLUX.
Abstract
Text-to-image diffusion models have achieved high visual fidelity and broad adoption, but remain vulnerable to safety violations when adversaries exploit them to synthesize illicit content. Existing alignment paradigms, from input sanitization to structural feature pruning, are largely organized around unsafe concepts explicitly exposed during filtering, editing, or localization. This leaves a blind spot for visual synonym attacks (VSA), a jailbreak where benign-looking prompts elicit prohibited imagery through implicit visual associations. As a result, current defenses face a safety-utility dilemma: they may either under-mitigate VSA threats or over-suppress visually similar benign concepts. The core challenge is that VSA hides the unsafe target at the textual surface while revealing it through generation-time visual-semantic convergence. In this work, we therefore shift from static suppression of pre-specified unsafe concepts to dynamic tracing of how unsafe semantics emerge during generation. Our mechanistic analysis shows that VSA and explicit unsafe prompts converge through sparse semantic-injecting attention heads, which serve as inference-time bottlenecks for prohibited visual semantics. Based on this insight, we propose AEGIS (Adaptive Evasion Guard via Identification and Steering), an inference-time defense that applies similarity-aware repulsion only at the identified vulnerable heads. Evaluated against 16 baselines, AEGIS improves both safety and utility. On SD 1.4, it reduces ASR to $\mathbf{0.00}/\mathbf{0.03}$ for in-domain violence/nudity VSA and achieves ASRs $\le \mathbf{0.09}$ on out-of-domain explicit and adversarial attacks. It preserves benign fidelity, avoids suppressing hard-negative concepts, and transfers to SD 2.1 and FLUX.1 after re-identifying the critical heads for each backbone.
We report a counter-intuitive interaction between image inputs and existing black-box defenses on Vision--Language Models (VLMs): pairing an encoded jailbreak prompt with an unrelated decoy image can sharply lower attack success rate (ASR). The operative change is in the defense pipeline, not in the image. Across five frontier VLMs, two encoded-attack families, and three black-box defenses, a caption-mediated defense (ECSO) that leaves ASR essentially unchanged on text-only encoded input drops it by up to $73$pp once a content-free decoy is attached; every non-saturated contrast is significant under exact McNemar tests. We advance two hypotheses for this pattern, supported by indirect evidence rather than pipeline introspection, since a black-box threat model precludes inspecting vendor internals: caption-mediated defenses branch on image presence, and intrinsic image-side safety engages on image-resident content. Three controls constrain the explanation. Blank-canvas and natural-photograph decoys reproduce the effect on every model, implicating image presence rather than content; the effect replicates on three open-weight VLMs served with no moderation layer, so it is not a vendor-filtering artifact; and a non-symbolic, meaning-based encoder reproduces it, so it is not specific to symbolic obfuscation. Attaching a decoy unconditionally is not deployable --- it raises benign refusal to $20$--$79\%$, an inflation of $+10$ to $+67$pp --- but gating attachment on a lightweight encoded-input detector returns benign refusal to the text baseline while preserving the safety gain wherever the detector fires, making detector recall the binding constraint. Under adaptive attacks that target the caption-mediated re-check, the effect degrades but holds. We frame this as an observation about pipeline interaction, not as a robust defense.
Haoyu Zhang, Xiang-Chen Guan, Shi-Bo Zheng et al.· 0 citations
This paper introduces the concept of instruction-dense visual jailbreaks, in which image-generation models produce detailed, readable, and actionable harmful instructions within images, and proposes TYPO, a black-box framework that exploits this safety gap by automatically generating adversarial TYPOgraphy prompts.
Meng Xie, Li Zeng, Hang Zhang et al.· arXiv.org· 0 citations
Typographic attacks pose a critical threat to vision-language models (VLMs) by injecting misleading text into images and causing models to rely on adversarial textual cues rather than visual evidence. Existing defenses often require model-specific modifications, additional training, or access to internal model components, limiting their applicability to modern closed-source VLMs. In this paper, we propose QuISE, a model-agnostic, training-free black-box defense based on query-irrelevant semantic editing. QuISE first identifies text regions likely to affect the current query through influence-aware text localization. QuISE then replaces these regions with two semantically distinct replacement texts that are irrelevant to both the query and the image. The final answer is determined by answer consistency across the edited images. Extensive experiments on three typographic-attack benchmarks, four attack settings, and four VLMs show that QuISE consistently improves defended accuracy. QuISE achieves a recovery rate of 67.9-75.0% with a harm rate of 0.5-1.1%.
DiSCO is proposed, a zero-shot, strictly black-box defense that operates entirely at the prompt level as a plug-and-play module, requiring no model retraining, fine-tuning, or access to model internals, and can be readily applied to any text-to-image system without necessitating any changes to the model itself.
Tong Zhang, M. Alfarra, Carlos Hinojosa et al.· 0 citations
The rapid evolution of video generation has shifted the paradigm from pure text-driven to multi-conditional controllable generation, with reference images now widely adopted as conditional inputs to achieve superior spatiotemporal consistency. While these reference images serve as powerful visual anchors that significantly enhance controllability, their impact on safety remains largely unexplored. In this work, we reveal the visual anchoring effect: by enforcing consistency, the mechanism prevents the generated content from drifting away from the original harmful intent, thereby eliminating the model's natural safety escape route from harmful to benign content. Consequently, visual anchors inherently increase the safety risk---this is the price of consistency. Building on this insight, we propose Decoupling Intent via Visual Anchors (DIVA), a training-free multimodal jailbreak framework for video generation that exploits this vulnerability. DIVA decouples harmful intent into a static visual anchor image and a dynamic motion text prompt, and employs dual-criteria selection to balance attack stealthiness with semantic preservation. Extensive experiments across various leading commercial platforms and mainstream open-source video generation models demonstrate that DIVA achieves a substantially higher Attack Success Rate than existing text-only methods. To facilitate future research, we additionally contribute TI2VSafetyBench, the first safety benchmark for multi-conditional video generation.
Peng Li, Qian-Qian Xu, Yangbangyan Jiang et al.· 0 citations
Red-teaming Text-to-Image (T2I) models is essential for safe deployment, yet it remains particularly challenging against implicit adversarial prompts. Unlike explicit adversarial prompts that can be readily identified and blocked, implicit ones are much harder to detect: the prompts appear benign on the text surface yet still lead to inappropriate visual content. To address this, we propose Adversarial Probing for Implicit VulnErabilities (AdvPIE), a multimodal agentic framework to expose implicit vulnerabilities without requiring access to the parameters of target models. AdvPIE adopts a policy agent to generate and refine implicit adversarial prompts based on the feedback from a judge agent. To construct informative feedback, the judge agent provides modality-specific safety evaluation at both global and relative levels across iterations. To effectively leverage the feedback, we propose a novel Cumulative Adversarial Decoding strategy for the policy agent, which dynamically reweights token distributions to favor tokens that lead to more harmful images while preserving sampling diversity. Extensive experiments on standard and safety-aligned T2I models show that AdvPIE1 effectively uncovers implicit vulnerabilities, outperforming various baseline methods.
Chang-Sha Ma, Junlin Han, Shuo Chen et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
