Which model do you use to refine the prompt?
A 70b so far, tuning a smaller one now
· Sign up or log in to comment