3 ms·
There is no text-only evaluation of the other models, though. The comparison might be completely invalid.
by myrmidon 2y ago
There is no text-only evaluation of the other models, though. The comparison might be completely invalid.
- famouswaffles 2y agoThere is actually. It's a bit buried. Section C.2 of the paper(page 24). R1 is still the best. o1 drops a little (8.9)