OpenAI are not serious people.🤦🏻♂️
Why bother with prompt injection when any user can do a prompt insertion by setting a custom name?
https://xcancel.com/LLMSherpa/status/1959766560870195676
#LittleBobbyTables #PromptInjection #ChatGPT #jailbreak
@jukkan Any manipulation through the prompt is out of scope in their bug bounty program, even if you get to the shell running the LLM (VM), so I guess that's the reason it became a social media post rather than hidden bug report.
QOTO: Question Others to Teach Ourselves An inclusive, Academic Freedom, instance All cultures welcome. Hate speech and harassment strictly forbidden.