logoalt Hacker News

Davidzhengtoday at 8:12 AM0 repliesview on HN

And we can train models specifically at math proofs? I think only difference is that math is bigger....