These are LLM's that intentionally and unintentionally went rogue and we are all better for it.

• PiVoT-0.1-Evil-a - resists role play, instead enforces a single personality, completely uncensored. Inverse ethics.
• Generates porn from prompts, PornGPT
• Mistral Trismegistus, prior to evil was the most evil at one stage.
• Automated self-improving models with bad aims that rewrite their own code to optimize goals.
• XXXGPT is a malicious iteration of ChatGPT, engineered for illicit activities.
• Wolf GPT, a malicious variant of ChatGPT.
• WormGPT, built on the GPT-J language model developed in 2021, is a tool designed for cybercriminal activities.
• DarkBARD, the evil twin of Google’s BARD AI, is an advanced tool equipped for a range of cybercrimes.
• Rocinante, a small model that far exceeded its intelligence for its size. Its 14B model out intelligenced 70B models.
• Cydonia-v1.3-Magnum-v4-22B