Per the API reference:
The maximum value for
logprobsis 5. If you need more than this, please contact us through our Help center and describe your use case.
I would like to access a matrix of logprobs for all classifiers in my fine-tuned model. I have about 30 classifiers, and many of the prompts fit into more than one of these.
Alternatively, should I create a distinct fine-tuned model for each classifier, trained on “X” vs. “Not X”? An example would be a survey comment about teamwork - “is about teamwork” or “is not about teamwork”.