用威胁类图像训练模型，让其自发产生安全意识。
Visual Self-Fulfilling Alignment: Shaping Safety-Oriented Personas via Threat-Related Images
- 在无安全标签情况下，用威胁类图像微调视觉语言模型。
- 攻击成功率下降，回应质量提升，过度拒绝现象缓解。
- 适合关注模型安全对齐但缺乏标注数据的研究者。
多模态大语言模型存在安全错位问题，视觉输入可能引发有害输出。现有方法依赖显式安全标签或对比数据，但威胁概念具象且可视，而安全概念如帮助性则抽象且无视觉对应。受涌现错位中自我实现机制启发，我们提出视觉自我实现对齐（VSFA）。VSFA在围绕威胁类图像构建的中性VQA任务上微调视觉语言模型，无需任何安全标签。通过反复接触威胁类视觉内容，模型内化警惕与谨慎的隐含语义，形成安全导向的人格。跨多个视觉语言模型和安全基准的实验表明，VSFA可降低攻击成功率，提升响应质量，并缓解过度拒绝，同时保持通用能力。本工作将自我实现机制从文本扩展至视觉模态，提供了一种无标签的视觉语言模型对齐方法。
原文摘要 · Abstract (English)
Multimodal large language models (MLLMs) face safety misalignment, where visual inputs enable harmful outputs. To address this, existing methods require explicit safety labels or contrastive data; yet, threat-related concepts are concrete and visually depictable, while safety concepts, like helpfulness, are abstract and lack visual referents. Inspired by the Self-Fulfilling mechanism underlying emergent misalignment, we propose Visual Self-Fulfilling Alignment (VSFA). VSFA fine-tunes vision-language models (VLMs) on neutral VQA tasks constructed around threat-related images, without any safety labels. Through repeated exposure to threat-related visual content, models internalize the implicit semantics of vigilance and caution, shaping safety-oriented personas. Experiments across multiple VLMs and safety benchmarks demonstrate that VSFA reduces the attack success rate, improves response quality, and mitigates over-refusal while preserving general capabilities. Our work extends the self-fulfilling mechanism from text to visual modalities, offering a label-free approach to VLMs alignment.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。