News for hackers. Stuff that matters.
Re:Legal models hallucinate in 1 out of 6 (or more) benchmar... by viraptor on Saturday June 01, 2024 @07:31AM (#40544821)
> we want them to tell us the truth, to give us the facts.

That's just one use case out of many. We also want it to tell stories, make guesses, come up with ideas, speculate, rephrase, ... We sometimes want facts. And sometimes it's more efficient to say "give me facts" and verify the answer then to find the facts yourself.

Re:Legal models hallucinate in 1 out of 6 (or more) benchmar... by surfingdino on Saturday June 01, 2024 @11:00AM (#40546282)
What if other sources of facts switch to confabulating LLMs? How will you be able to tell facts from made up information?
Re:Legal models hallucinate in 1 out of 6 (or more) benchmar... by mercer on Saturday June 01, 2024 @12:51PM (#40547272)
how do you do that now?