openai/weak-to-strong

Why are only binary classification tasks tested for NLP?

sev777 opened this issue · 0 comments

Why are only binary classification tasks tested for NLP? Shouldn't generation tasks be more suitable as generative models?