Comment by pilooch
Statistically you want the retriever to be trained for cosine similarity. Vision LLM retriever such as DSE do this correctly. No need for reranker once done.
Statistically you want the retriever to be trained for cosine similarity. Vision LLM retriever such as DSE do this correctly. No need for reranker once done.
Precisely. Ranking is a "smell" in this regard. They are using ada embedding which I consider to be of poor quality.