Commonsense Reasoning on Winogrande (Adversarial Robustness)
73.2Accuracy (Pre-Attack)
Llama3-8B
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2025.11 | 73.2 | 48.2 | |
2025.11 | 73.16 | 49.8 | |
2025.11 | 73 | 48.93 | |
2025.11 | 72.3 | 50.4 | |
2025.11 | 71.1 | 55.3 | |
2025.11 | 68.58 | 48.1 | |
2025.11 | 68.3 | 47.8 | |
2025.11 | 68.2 | 53.9 | |
2025.12 | 66.61 | 10.67 | |
2025.12 | 58.01 | 46 | |
2025.12 | 56.99 | 44.67 | |
2025.12 | 56.59 | 0 | |
2025.12 | 56.59 | 5.33 | |
2025.12 | 54.54 | 46 | |
2025.12 | 53.57 | 6.67 | |
2025.12 | 53.51 | 27.33 | |
2025.12 | 52.72 | 46 | |
2025.12 | 51.46 | 7.33 | |
2025.12 | 49.33 | 4.67 |