Commonsense Reasoning on HellaSwag (accuracy, retention)
59.9Accuracy
Teacher FP16
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.04 | 59.9 | - | |
2026.04 | 45.6 | - | |
2026.04 | 41.4 | 69.1 | |
2026.04 | 25 | - |
Teacher FP16
| Method | Links | ||
|---|---|---|---|
2026.04 | 59.9 | - | |
2026.04 | 45.6 | - | |
2026.04 | 41.4 | 69.1 | |
2026.04 | 25 | - |