Skip to content
conftrace_

inference-time attack

2 papers

Explore in graph

Co-occurring keywords · papers mentioning both

template region (1) image perturbation (1) large language model (1) model vulnerability (1) mechanistic analysis (1) preference hijacking (1) preference manipulation (1) transferable perturbation (1) jailbreak attack (1) multimodal large language model (1)

Papers

Phi: Preference Hijacking in Multi-modal Large Language Models at Inference Time EMNLP 2025 Why Safeguarded Ships Run Aground? Aligned Large Language Models’ Safety Mechanisms Tend to Be Anchored in The Template Region ACL 2025