based on the contents of the BIRD benchmark, this is a HUGE self report that this team leaked the test data / overfit to the answer set
that or they're p-hacking
those questions are extremely "up to interpretation"
显示更多
🚀 Introducing Gemini-SQL2, our breakthrough text-to-SQL capability powered by Gemini 3.1 Pro! We've achieved state-of-the-art results on the highly competitive BIRD benchmark, translating natural language into execution-ready SQL queries. 🧵👇
显示更多
