Prompt Privacy from LLMs

2 points | by napping_penguin an hour ago

1 comments

  • napping_penguin an hour ago

    I recently came across a really interesting piece of privacy technology. Suppose you have a model M and a prompt P. The technique allows you to create an obfuscated prompt Q such that:

    - M(Q) is nearly the same as M(P)

    - P is hard to reverse engineer from Q

    As a applied crypto researcher, this feels like an "ML-based homomorphic encryption". Works with any model (that supports prompt_embeds) without changing anything on the model side. Very cool indeed.

    Credit note: This method was invented by Protopia Labs and I don't have any affiliation there.