Category: Benchmarks

This category contains 1 page.

• MMLU · Massive Multitask Language Understanding, a 57-subject multiple-choice knowledge benchmark introduced in 2020 and long the standard headline eval for language models.