Why Pinecone vector db result is not perfect for retrieval on complex query?

Hi, I’ve embedded the Wikipedia pages of 50 United States presidents using a Hugging Face model and stored those embeddings on Pinecone. I asked a simple question: ‘Please provide me with the name of the college that most U.S. presidents attended, and also give me a list of those presidents.’ However, it’s giving me the right similar vector embeddings. Can anybody help me with how to solve this type of problem?

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.

@ai1 I’m not sure I understand the issue. When you say it’s giving you the right similar vector embeddings, do you mean you’re getting the results you expect?

Can you give an example of the queries you’re running, the results you expect to get, and the results you’re actually getting?