1
From Prompt Risk to Response Risk: Paired Analysis of Safety Behavior of Large Language Model
大模型安全新视角：从提示风险到响应风险的配对分析，揭示LLM安全行为的深层机制。
arXiv:2604.26052v3 Announce Type: replace Abstract: Safety evaluations of large language models (LLMs) typically report binary outcomes, i.e. attack s…