Replies: 1 comment
-
An additional concern in this respect: RAG models are very useful to extract relevant information from a lengthy documentation. However they can only work with limited context sizes. I have seen hallucinations in RAG models trying to answer Colvars questions, that I suspect were due to the doc not fitting in the model context. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I think the main use case of our doc is online in HTML form. The single HTML page is becoming large and unwieldy. I've experimented with tex4ht to generate separate pages with a side frame containing a TOC. This roughly works, but not very well because the Latex code + building toolchain is complex and somewhat fragile.
I'd like to know if there is demand for PDF docs. I suspect we could just drop them and no one would bat an eye. This would remove the need for supporting different cases with latex macros, and let us focus on the main stake, the HTML.
Beta Was this translation helpful? Give feedback.
All reactions