logoalt Hacker News

dwaltrip • today at 12:21 AM • 0 replies • view on HN

RL is part of the model’s training. It changes the weights.

What distinction are you drawing?