Please do and thanks in advance for any insights you can provide -- it would be great to understand any benchmarking improvement with ada-002 from your previous findings, and whether you tested the specific OpenAI text-search-*-{query,doc} models as a comparison for large document search.