alexdzm·anno scorso·discussKeen to hear more about this benchmark. Is it representative of chat-to-document style usecases with big docs?