跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.23206 · v1

Reading Scene States: Uncovering Safety Failures in Vision-Language Models

首次提交 2026-11-18 02:48
版本更新 2026-11-18 02:48
雷达观测 2026-11-18 12:34
arXiv ↗PDF ↗

摘要

Vision-language models (VLMs) can provide harmful assistance by interpreting visual events, even without an explicit request for the underlying harmful objective. We identify this \textbf{scene-state safety vulnerability} and introduce \textbf{SceneSafetyBench}, comprising 375 image--question pairs spanning Initial, Medium, and Ultimate scene states and five visual presentations. Across eight VLMs, scene-state inputs elicit harmful assistance with attack success rates (ASRs) of 45.6--72.3\%. Comparative analysis across three VLMs shows that scene-state inputs most closely resemble QR-related images in their layer-wise safety-related signal profiles, while exhibiting lower average late-layer readouts and distinct signal dynamics during generation. To turn this vulnerability into a systematic attack, we propose \textbf{S3TA}, which translates harmful objectives into scene states, queries image--question pairs independently, and assembles the responses. On seven MM-SafetyBench categories across eight open-weight and commercial VLMs, S3TA achieves a model-average ASR of \textbf{71.2\%, exceeding the strongest baseline by 6.1 percentage points}. Text-side filters and visual moderation provide limited protection. These findings highlight the need to safeguard the assistance models infer through scene-grounded reasoning, beyond screening input content.

作者备注:VLM safety, multimodal jailbreaks, and scene-state reasoning

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-18 02:48 · 最新观测版本

Reading Scene States: Uncovering Safety Failures in Vision-Language Models

Vision-language models (VLMs) can provide harmful assistance by interpreting visual events, even without an explicit request for the underlying harmful objective. We identify this \textbf{scene-state safety vulnerability} and introduce \textbf{SceneSafetyBench}, comprising 375 image--question pairs spanning Initial, Medium, and Ultimate scene states and five visual presentations. Across eight VLMs, scene-state inputs elicit harmful assistance with attack success rates (ASRs) of 45.6--72.3\%. Comparative analysis across three VLMs shows that scene-state inputs most closely resemble QR-related images in their layer-wise safety-related signal profiles, while exhibiting lower average late-layer readouts and distinct signal dynamics during generation. To turn this vulnerability into a systematic attack, we propose \textbf{S3TA}, which translates harmful objectives into scene states, queries image--question pairs independently, and assembles the responses. On seven MM-SafetyBench categories across eight open-weight and commercial VLMs, S3TA achieves a model-average ASR of \textbf{71.2\%, exceeding the strongest baseline by 6.1 percentage points}. Text-side filters and visual moderation provide limited protection. These findings highlight the need to safeguard the assistance models infer through scene-grounded reasoning, beyond screening input content.

查看此版本解读 →查看此版本原文 ↗