Does using a fine tuned GPT3 further fine tunes the model?

If I fine tuned one of the GPT-3 models with my data, and after that I used it with new prompts for completion, will the new prompts also add to the fine tuning, or do I need to explicitly fine tune the model with the data further, using fine tuning endpoint.

i.e. Say I trained my model to classify certain text into happy or sad.Once trained, I prompted my model with a brand new text and say it completed it properly. Now if I prompted my model again saying that it predicted right. Will this interaction with the model further fine tune the model?

1 Like


Prompting interactions do not modify the model—nor would you want them to… Imagine all your users being able to fine tune the model you’re making available…

1 Like