[Shitty-kitty]

rag basically gives the llm a bunch of documents to search thru for the answer. What it doesn't do is make the algorithm any better. pre-training and fine-tunning improve the llm abaility to reason about your task.