MMLU
MMLU (Massive Multitask Language Understanding) is a benchmark test launched by the University of California, Berkeley in September 2020, aiming to comprehensively evaluate large language models' multitask understanding across 57 different domains.