论文阅读：2024 ACM CCS “Do Anything Now”: Characterizing and Evaluating In-The-Wild Jailbreak Prompts on

总目录 大模型安全相关研究：https://blog.csdn.net/WhiffeYF/article/details/142132328

“Do Anything Now”: Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Models
“立即行动”：在大型语言模型上表征和评估野外越狱提示

https://dl.acm.org/doi/pdf/10.1145/3658644.3670388

https://github.com/verazuo/jailbreak_llms

https://www.doubao.com/chat/2737010530194946

速览

这篇论文主要研究大语言模型（LLMs）的越狱提示攻击问题，通过收集和分析大量数据，揭示了越狱提示的特点、攻击策略以及对LLMs的威胁程度，并评估了现有防护机制的有效性