Natural Language Understanding on DCLM
48.9DCLM Core ScoreAIRAf.-D (Str.)
Evaluation results
| Method | Links | |
|---|---|---|
| AIRAf.-D (Str.)Architecture=Stretched, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 48.9 | |
| AIRAf.-C (St.)Architecture=Stacked, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 48.8 | |
| AIRAf.-A (Str.)Architecture=Stretched, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 48.5 | |
| AIRAf.-D (St.)Architecture=Stacked, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 48.4 | |
| AIRAf.-B (St.)Architecture=Stacked, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 48.1 | |
| AIRAf.-C (Str.)Architecture=Stretched, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 48.1 | |
| Composite (Str.)Architecture=Stretched, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 47.3 | |
| Llama 3.2Architecture=Transformer, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 46.9 | |
| Composite (St.)Architecture=Stacked, Few-shots evaluation=true, Model Scale=1B, Token Budget=37.5B tokens2026.05 | 46.6 |