豆包大模型团队Open SourceBenchmark测试集SuperGPQA

Jinshi Data News on March 4th, according to the official WeChat message of the Doubaoda model team, recently, the Doubaoda model team Open Source Super GPQA, a comprehensive and highly distinctive knowledge reasoning Benchmark test in the field. According to the introduction, the dataset has constructed an evaluation system covering 285 graduate-level disciplines, including 26,529 professional questions, not only covering mainstream disciplines, but also including long-tail disciplines such as light industry, agriculture, and service science, demonstrating the breadth of coverage of comprehensive disciplines, filling the gap in the long-tail knowledge evaluation field.