A professional evaluation benchmark from Fudan University’s NLP Lab designed to measure the performance and reliability of large language models.