Psst wanna jailbreak ChatGPT? malicious prompts for sale

Turns out it's pretty easy to make the model jump its own guardrails

Read more here: External Link