Hi i am new to this field and my goal is to train (fine tuning) LLM model .
i have read and understand i believe the basic for create an Alpaca form dataset (csv file or excel with 3 columns instruction, input, output)
My Question is how to construct the database if i want my LLM to prompt me if an answer depends from something else.
Let’s say that i want to fine tuning it for assist of vacuum cleaner usage .
we have 3 brands of vacuums (Brand01, Brand02, Brand03)
- Brand01 has a turn switch to power on,
- Brand02 has a toggle switch
- Brand03 has a push on / off switch
How would like my LLM work :
User - Question : why the vacuum cleaner not working
LLM - Answer : Check if the vacuum cleaner cable is on socket, is it?
User - Question : Yes
LLM - Answer : What Brand of vacuum cleaner do you own?
User - Question : Brand03
LLM - Answer : Please Check if the Push on/ off switch is turned on
Something like that ,
Any Help - Any suggestions will be welcomed
Thanks in Advance