logoalt Hacker News

nomeltoday at 5:33 PM0 repliesview on HN

Hasn't it been repeatedly shown that many small models perform worse than a large model of the same total parameters?