Fine-tuning LLM for expected behaviour

API
,
1

Hello,

I am working on fine-tuning and have gone through the resource available on openAI website. I have a query regarding the fine-tuning the behaviour of the LLM with completion APIs. If the system is expected to have several identified behaviours, is it possible to train the LLMs on these identified behaviours at once and expect it to behave in a particular way using prompt?

Regards

2

am working on fine-tuning and have gone through the resource available on openAI website. I have a query regarding the fine-tuning the behaviour of the LLM with completion APIs. If the system is expected to have several identified behaviours, is it possible to train the LLMs on these identified behaviours at once and expect it to behave in a particular way using prompt?

Yes, I recomend reading over the documentation on fine tuning. depending on what you need, there is multi turn and single turn fine tuning

4

Thanks for the reply will check multi-turn.

5

cool, feel free to mark my comment as a solution, that’ll help future community visitors searching and landing on this topic. oh, and welcome to the community!