Re:Try Deep Think in the Gemini App
by Davidzheng on Friday August 01, 2025 @08:13PM
(#44763814)
but if the less strong model has low false positives you can just route them in order of strength
[ Parent
| on HN ]
That's a very big "if".
Re:Try Deep Think in the Gemini App
by Davidzheng on Saturday August 02, 2025 @08:40PM
(#44773078)
I very much think a lot of false positives is caused by incentives to maximize score where incorrect answers are not penalized