I am trying to fine-tune a davinci model for experience sharing which is not working very well so far.
I found the section “More prompt advice” on OpenAI CookBook with wonderful advices. Tested on OpenAI playground page these advices have really improved the situation.
But to repeat all the advices into every prompts and completions in the jsonl training data lines is not realistic. So is there a better way?
Thank you very much!!
Welcome to the community!
I think the only way is to redo your dataset, unfortunately. I’ll be watching the thread to see if anyone else has ideas, though.
Thank you very much!
When redo dataset still need a hit about how to do advices from More prompt advice like this one:
- If you want it to say “I don’t know” when it doesn’t know the answer, tell it ‘Say “I don’t know” if you do not know the answer.’
Sounds like training data jsonl has a field other than ‘prompt’ and ‘comletion’ probably be ‘instruction’:''Say ‘I don’t know’ if you do not know the answer."?
No, I think it means to include that in the prompt section.
Feel free to share the prompt in the “Prompt Help” forum, and we can take a stab at improving it for you.
Jump this thread to “Prompt assistant” forum