Re:Elon Musk's Grok praises Hitler, shares antisemitic trope...
by salawat on Wednesday July 09, 2025 @08:22AM
(#44509162)
You misunderstand. There is no Quality Control you can do for these things. They're ill-conditioned function imitators. Small changes in input, yield massive changes in output. You can't QC that. You can only clean up after it.
[ Parent
| on HN ]
Re:Elon Musk's Grok praises Hitler, shares antisemitic trope...
by jonnycomputer on Wednesday July 09, 2025 @10:21AM
(#44510406)
Of course you can. Its not as easy as software testing, but you certainly can run it through a gamut of prompts intended to provoke it to act in ways that you find unacceptable. I mean there is a whole field studying and applying AI alignment.
Re:Elon Musk's Grok praises Hitler, shares antisemitic trope...
by electrondood on Wednesday July 09, 2025 @12:13PM
(#44511791)
That's completely false. You run evals.
edit: you're supposed to run evals.