DeepSeek V2 on BBH (BIG-Bench Hard): Score Breakdown

DeepSeek V2 × BBH (BIG-Bench Hard)

Puan Dökümü

Model SıralamasıPuan
#105 / 20570.5

Performans Analizi

DeepSeek V2 achieves 70.5 on BBH (BIG-Bench Hard).