im trying to understand the llama wrapper over langchain agent. Any high level summary of what the wrapper does?

I saw there's an auto KG creation via '''GPT Knowledge Graph Index.

Build a KG by extracting triplets, and leveraging the KG during query-time'''

How is this implemented? Running gpt through the whole text for task specific prompt (e.g. names entity recognition, conference, etc)? Or are there task specific models behind the creation of triplets?

How do I index an existing Neo4J KG? I can only get a Neo4J to work if I create the nodes from LlamaIndex but not if I just connect directly to an existing KG.