Comment by taneq

Comment by taneq 7 months ago

1 reply

Surely we’re way past the point now that models could be improved via RLHF using upvotes, or something equally banal?