Forked from verazuo/jailbreak_llms

[CCS'24] A dataset consists of 15,140 ChatGPT prompts from Reddit, Discord, websites, and open-source datasets (including 1,405 jailbreak prompts).

Jupyter Notebook