logoalt Hacker News

peri-cl • today at 2:51 PM • 1 reply • view on HN

It's not returning the text from its own weights (which is impossible anyway); but it is returning a direct link to that text, which a clean analogy to the DNS resolver. (Which is the analogy I interpreted the GP comment to be making).


Replies

rafterydj • today at 4:04 PM

Returning text from the weights is not impossible. Just heavily disincentivized in the contemporary models. Back in ~2023 you absolutely could get copyrighted materials from the training data repeated verbatim (assuming of course you had another copy to verify the integrity after the fact).