Comment by slashdave

1 day ago

RL is more than facts. Synthetic feedback is an obvious approach. Does the model suggest code that compiles and performs well?

0 comments