Is there a topic for fine-tuning discussion?

I have a requirement these days to put our data to GPT-3.5 (actually I want to use get-4, I don’t have permission to do it. Please let me know if you have any way to get it on API), and then get the fine-tuning model. After I got the new model, it is not good as I think. I know there are some parameters when create fine-tuning, suck as hyperparameters. Do you have any idea about the parameter and how the parameter affects the fine-tuning result?

Is there a discuss to join to solve the detail problem like this?

This might be of assistance: