When you fine tune a model, you are making its responses more granular. So, based on how you have fine tuned it, you will get a detailed response.
From my view, the fine tuned model is working as expected.
The way to fix this would be to fix how your prompt/completions are created in the first place. You should ideally add more content to your completions to give the model an indication that you want more data back. Once you have fine tuned it, usually changing temp does nothing to it.