Re:GPT-5
by bigyabai on Thursday August 07, 2025 @01:55PM
(#44827955)
6. (Occam's Razor) It just didn't perform that well in trials for that specific eval.
[ Parent
| on HN ]
Re:GPT-5
by renewiltord on Thursday August 07, 2025 @02:04PM
(#44828112)
That is obviously wrong since the numbers are right but the graph is wrong and you can see it correct on the website…