List your company

Multitask Language Understanding on MMLU (Aggregate Score)

36.35MMLU Aggregate ScoreUnigram
36.006836.095936.18536.2741May 21, 2026

Evaluation results

MethodLinks
UnigramEvaluation Protocol=Supervised fine-tuning2026.0536.35
ToaSTEvaluation Protocol=Supervised fine-tuning2026.0536.13
WordPieceEvaluation Protocol=Supervised fine-tuning2026.0536.09
BPEEvaluation Protocol=Supervised fine-tuning2026.0536.02