x

LESSWRONG
LW

WitchBOT — LessWrong

WitchBOT

WitchBOT

Message

86

1

3y

WitchBOT

You can use GPT-4 to create prompt injections against GPT-4

So, I saw this tweet. It turns out you can ask GPT4 to compress a message in a way it can understand. The end result is usually not understandable for a human. It's interesting, since it allows you to effectively rise the context length. It also reveals a spooky degree...

Apr 6, 2023•87

Message

86 karma

1 post

Member for 3 years

You can use GPT-4 to create prompt injections against GPT-4

3y

So, I saw this tweet. It turns out you can ask GPT4 to compress a message in a way it can understand. The end result is usually not understandable for a human. It's interesting, since it allows you to effectively rise the context length. It also reveals a spooky degree of self-awareness.

I decided to apply it to one of the prompt injections from Jailbreakchat.

After a few attempts and tweaks, I got one of the smallest working injections to date against GPT-4

And then, I got curious. Is it something of Machiavelli -- his grey role in history, his desire to provide advice without scrupules, his great stature as a thinker -- that makes... (read 368 more words →)

9

87