where is the prompt template that I can edit for when I call query on a query engine?

Is there a benchmark somewhere of retrieval augmented generation performance vs trying to get the model to learn this information via fine-tuning?

@kapa.ai

Let's say I had a List index, vector index, and a kw index.

Is there a way to query from all of the indexes and use information from all three indices to formulate an answer?

@kapa.ai does llama-index work with Spotify's annoy library?

@kapa.ai Can you explain a few good usecases for Composable Indices within the llama-index library?

@kapa.ai how do I convert a Document into a Node in llama-index/gpt_index