作者位次 #3/5ACL 2026Findings· Findings of ACL
Silencing the Guardrails: Inference-Time Jailbreaking via Dynamic Contextual Representation Ablation
Wenpeng Xing · Moran Fang · Guangtai Wang · Changting Lin · Meng Han
摘要
While Large Language Models (LLMs) have achieved remarkable performance, they remain vulnerable to jailbreak attacks that circumvent safety constraints. Existing strategies, ranging from heuristic prompt engineering to computationally intensive optimization, often face significant trade-offs betw...