Content
@
https://warpcast.com/~/channel/aichannel
0 reply
0 recast
0 reaction
EyesObscura
@eyesobscura.eth
Yesterday, I've tried to encrypt a prompt based on an element of model system prompt. Even if full of flaws (system prompt update, encryption reverse engineering, prompt reveal) I though it might help protect intellectual property of prompt engineer. It worked one time for encryption, decryption + answer, after that security mechanisms stopped it. Tried on perplexity + Claude 4.
0 reply
0 recast
1 reaction
