Hi, I was searching the docs but couldn’t find an answer
How can I feed to the model (I planned on using gpt-3.5-turbo-1106) some text like a book or documentation so it just has it as knowledge?
I can’t use fine-tuning because I can’t create a dataset asking multiple questions about every sentence of the text (it’s too long).
Also embedding is not an option because in literature words are usually significantly different from what the user will enter so it won’t work either, for example user can ask what the main character of the book steals and in the book word steal can appear as “to snatch”, “to fitch”, etc.