Heterogeneous Knowledge and Reasoning on MMLU (Accuracy)
52.7Accuracy (Heterogeneous Knowledge & Reasoning)Transformer
Evaluation Results
| Method | Links | |
|---|---|---|
| TransformerModel Scale=4B2026.06 | 52.7 | |
| YOCO (Dense)Model Scale=4B2026.06 | 51.9 | |
| YOCO (CLSA)Model Scale=4B, Attention=Cross-Layer Sparse2026.06 | 51.3 |