When OpenAI makes an improved model, do they just correct its “wrong answers” till it spits out more correct-ness?

Question

When OpenAI makes an improved model, do they just correct its “wrong answers” till it spits out more correct-ness?

In: Technology

Answer 1

[removed]