Hello,
I use chatGPT API to decide whether 2 questions in specific context are actually the same question, just written differently. (Questions are always different). Turns out chatGPT can decide whether 2 questions are the same very well.
My question is if fine tuning a model would improve how well chatGPT decides whether they are the same question or not. Or is fine tuning just to ensure output is in a specific format?