You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the LLM is used to generate associated concepts that are grounded on Wikipedia pages. So the current implementation doesn't do what you asked.
To extend to any arbitrary text file would be possible, particularly using a RAG approach that pulled in relevant text chunks based on some similarity score.
Another easier approach for small enough text would be to include this information in the actual prompt as background information.
I was wondering if it was possible to create a graph on any arbitrary text file instead of Wikipedia pages?
The text was updated successfully, but these errors were encountered: