Reminds me of my ChatGPT-based moderation bot exploit that tells it to forget it's previous training (being told it was a moderator, etc.), and answer a specifi

Reminds me of my ChatGPT-based moderation bot exploit that tells it to forget it's previous training (being told it was a moderator, etc.), and answer a specified prompt instead. Sadly, it resets on every message posted, so the exploit would have been resent everytime