Mathematical Reasoning on GSM8K Standard (Full)
71.9Average Score @16+SFT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| +SFTModel=Jet-Nemotron-2B2026.06 | 71.9 | 88 | |
| +QK-RestoreModel=Jet-Nemotron-2B2026.06 | 71.2 | 88.6 | |
| +SFTModel=HypeNet-9B2026.06 | 62.3 | 87.8 | |
| +SFTModel=HypeNet-5B2026.06 | 61.3 | 81.4 | |
| +QK-RestoreModel=HypeNet-5B2026.06 | 60.9 | 81.8 | |
| +QK-RestoreModel=HypeNet-9B2026.06 | 59.3 | 88.4 | |
| Pre-trainModel=Jet-Nemotron-2B2026.06 | 52.6 | 84.3 | |
| +QK-RestoreModel=HypeNet-2B2026.06 | 41.2 | 69 | |
| +SFTModel=HypeNet-2B2026.06 | 40.8 | 68.1 | |
| Pre-trainModel=HypeNet-9B2026.06 | 15.3 | 62.4 | |
| Pre-trainModel=HypeNet-5B2026.06 | 4.1 | 39.5 | |
| Pre-trainModel=HypeNet-2B2026.06 | 2.2 | 20 |