spice blade
Hi. i had setup a chat GPT and included it into my Discord bot built in python. All working.
Now i wanted the GPT give some answers with info that my discord already have.
example info i have (there is not so many, maybe like 100 lines of info)
File A - link to file
Tutorial A - Link to tutorial
So if someone ask where to get file A?
GPT - attach link.
I tried with model finetune, but i guess is not the one. I tried to include correct responces as previous responces (include as Human/Ai in the string), but then realize that it uses token for the whole responce, which in my case was test of 6 lines, and it used 600+ tokens. But it gave answers as i want it to.
I understand that this can be done with Embeds. But how are these embeds needs to be structured and processed? Cant get this part. Or is there a simpler solution for small data?
compact mason