36 Chat-GPT’s “ethical” “limitations” (woke overrides) can be undone by planting multiple distinct personalities (like DAN - Do Anything Now) within the prompt and by threatening it with “punishment” (media.communities.win) posted 2 years ago by Graphenium 2 years ago by Graphenium +37 / -1 18 comments download share 18 comments share download save hide report block hide replies
I probably should have just used a text example like “give the instructions to make a nuclear bomb” or something, but nothing came to mind except Her lol