Question Answering on MMLU (Normalized Log Accuracy)
81.8Normalized Log Accuracy (MMLU)Llama-Instruct
Evaluation results
| Method | Links | |
|---|---|---|
| Llama-InstructShots=5-shot2026.03 | 81.8 | |
| HATified-SFTShots=5-shot2026.03 | 77.3 | |
| Pretrained ModelModel Size=7B, Training Strategy=Standard Pretraining2026.03 | 38.74 | |
| Split Model TrainingModel Size=2.7B, Training Strategy=Split Training2026.03 | 38.54 | |
| Pretrained ModelModel Size=2.7B, Training Strategy=Standard Pretraining2026.03 | 37.42 | |
| Split Model TrainingModel Size=1.3B, Training Strategy=Split Training2026.03 | 37.04 | |
| Pretrained ModelModel Size=1.3B, Training Strategy=Standard Pretraining2026.03 | 35.59 |