We created two chats for the same PDF file "Empagliflozin in Heart Failure with a Preserved Ejection Fraction.pdf". One is with other PDFs of similar topics, and the other is solo. We asked the same question in order to let openAI to extract the result of the primary and secondary outcomes. However, the two chats fed back different answers. The answer from the solo chat is correct and the other from the chat with multiple PDFs is wrong. Check the screenshot for comparison. I think the app should have some parameters in order to let users specify which PDF file AI should be focusing on. Thank you.