[–] robrenaud 2y ago ↗ As models improve, their representations converge. This even happens across modality for vision and text.
1 comment
[ 5.8 ms ] story [ 11.8 ms ] thread