> it IS exposed through API as logprobs in the ChatGPT api Sure but they often are not necessarily easily interpretable or reliable. You can use it to compare

> it IS exposed through API as logprobs in the ChatGPT api Sure but they often are not necessarily easily interpretable or reliable. You can use it to compare a model’s confidence of several different answers to the same question but anything else gets complicated and not necessarily that useful.