Translation in progress Some entries still include text in the original language.
Oct 12 · Monday
Defenses and guardrails · Papers
Found 4 papers提出 Safety-Aware Pruning，缓解视觉 token 剪枝放大的多模态越狱
- Published
- Evaluated models
- LLaVA-1.5-7B
Abstract剪枝去掉背景 token 后注意力塌向恶意前景。
这篇工作评估视觉 token 剪枝如何改变 VLM 在多模态越狱下的行为，并给出推理时防御 SAP。
提出 AEGIS，用中层内部信号选择性激活后层适配器拦截音频越狱
- Published
- Evaluated models
- Qwen2-Audio-7B, VITA-1.5, Ultravox v0.5 and others6
提出 SPARK，在预填充阶段修复多模态 KV 记忆以防御视觉语言模型越狱
- Published
- Evaluated models
- LLaVA-OneVision-7B, Chameleon-7B, Qwen2-VL-7B and others4
提出推理时防御 SafeThink，监视思维链并在低于阈值时注入转向前缀
- Published
- Evaluated models
- R1-Onevision-7B, OpenVLThinker-7B, VLAA-Thinker-7B and others6
You have reached the end