[–] robrenaud 2y ago ↗ As models improve, their representations converge. This even happens across modality for vision and text.
1 comment
[ 4.5 ms ] story [ 16.3 ms ] thread