Follow up question. How does GPTSimpleVectorIndex differ from openai’s embeddings?
My understanding is that the difference is that you use a node/graph structure that gives better relations between the embeddings vs openai’s embeddings that just clusters them. Is that correct?
