Commonsense Reasoning on HellaSwag (Accuracy and Pruning Analysis)
60Accuracy
Unpruned
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.04 | 60 | 0 | |
2026.04 | 37 | 0 | |
2026.04 | 35 | 19.6 | |
2026.04 | 35 | 0 | |
2026.04 | 34 | 0 | |
2026.04 | 27 | 41.7 | |
2026.04 | 26 | 24.3 | |
2026.04 | 22 | 34.6 |