Hi,
I need help with the Information extractor node, particularly with the LM tool.
I need to reed and retrieve information from a file. The output from the LM is always “see section 6.1” (an existing section in the document with the information I need), but since the whole point of the flow is not to go trough all the document to retrieve those information, how can I fix it? is there a good prompt I could use to avoid this annoying output?
There is a function called RAG, where the information is transformed into metadata and placed in a database, and it becomes much smaller when compared in rows. It can be done both with the postgres node and with the supabase VECTOR STORE node. I believe it can help you.
This RAG function is a set of nodes and uses LLM. It’s a bit difficult to explain here, but you can search for how to do RAG on YT. A short video will help you.
Thank you very much for the answer, I thought that the extractor node would work exactly as a RAG
(once I manage to fix the flow to run correctly I’ll mark your answer as the solution)