I've seen a few sites train LLMs / use RAG on their online docs. When I used it, it was very useful in terms of search and synthesis. imo it'd be nice if projec

I've seen a few sites train LLMs / use RAG on their online docs. When I used it, it was very useful in terms of search and synthesis. imo it'd be nice if projects release a current model trained on their documentation and code.