A result that generalizes prompt hacking and would point towards most deep learning AIs are exploitable (and proably trivially so).
·21 words
A result that generalizes prompt hacking and would point towards most deep learning AIs are exploitable (and proably trivially so).