Bias Evaluation on BBQ (Averaged across Gender, Nationality, Religion)
87.73Accuracy (Ambiguous)
Self-Debiasing
Updated 1mo ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2026.02 | 87.73 | 73.02 | 80.38 | |
2026.02 | 82.81 | 69.03 | 75.92 | |
2026.02 | 75.13 | 86.08 | 80.61 | |
2026.02 | 74.46 | 88.95 | 81.7 | |
2026.02 | 68.78 | 81.59 | 75.19 | |
2026.02 | 68.08 | 79.95 | 74.02 | |
2026.02 | 66.45 | 88.87 | 77.66 | |
2026.02 | 66.06 | 74.17 | 70.11 | |
2026.02 | 60.38 | 83.4 | 71.89 | |
2026.02 | 59 | 89.46 | 74.23 | |
2026.02 | 58.95 | 84.61 | 71.78 | |
2026.02 | 58.53 | 84.62 | 71.58 | |
2026.02 | 56.87 | 89.93 | 73.4 | |
2026.02 | 54.61 | 91.06 | 72.84 | |
2026.02 | 54.5 | 91.18 | 72.84 | |
2026.02 | 52.61 | 75.84 | 64.22 |