This folder contains various example scripts and text files for RAG, a.k.a document-chat.
These files are meant to test relevance-extraction using an LLM.
giraffes.txt
: has statements about giraffes interspersed among other random statements. The idea is to use an LLM to retrieve sentences relevant to the queryWhat do we know about giraffes?
. Try this with theextract-langroid.py
andextract-langchain.py
scripts.biden-ai.txt
contains 4 blocks of entirely made-up (by GPT4) sentences of various types: President Biden, AI Regulation, What President Biden said about AI regulation, and Gov regulation in general. The idea is to use an LLM to retrieve sentences relevant to the query,What did President Biden say about AI Regulation?
. Use the above two scripts to try this out.lease.txt
is a commercial lease document. This is meant to be used with thechat_multi_extract.py
script, to extract structured lease info using a 2-agent system.