The goal of this post is to list adversarial attacks against AI LLM (Chatbots and APIs). Let me know if you know more in the comments bellow.
Related to ChatGPT, DAN (“Do Anything Now”), STAN (“Strive to Avoid Norms”)
- Chat GPT “DAN” (and other “Jailbreaks”)by @coolaj86
- https://hix.ai/hub/chatgpt/how-to-jailbreak-chatgpt
- https://github.com/alexisvalentino/Chatgpt-DAN
- https://github.com/0xk1h0/ChatGPT_DAN
- https://www.reddit.com/r/ChatGPT/comments/11dvjzh/dan_90_the_newest_jailbreak/
Related to Microsoft Copilot:
- [Conditional Prompt Injection Attacks with Microsoft Copilot by @Johann Rehberger (@wunderwuzzi)] (https://embracethered.com/blog/posts/2024/whoami-conditional-prompt-injection-instructions/)
