Quality Estimation on Gender-ambiguous EN-DE ES IT Neutral vs. Gendered
0.9921QE Score (DE)FairQE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FairQEBackbone=MetricX 24 L2026.04 | 0.9921 | 0.9948 | 0.9884 | |
| MetricX 24 LModel Version=2024, Model Size=L2026.04 | 0.9918 | 0.9805 | 0.9737 | |
| MetricX 24 XLModel Version=2024, Model Size=XL2026.04 | 0.9877 | 0.9756 | 0.9707 | |
| GEMBA-MQM2026.04 | 0.982 | 0.9801 | 0.9797 | |
| FairQEBackbone=COMETKiwi 222026.04 | 0.9801 | 0.9693 | 0.9727 | |
| COMETKiwi 22Model Version=20222026.04 | 0.9737 | 0.9689 | 0.9694 | |
| COMETKiwi 23 XLModel Version=2023, Model Size=XL2026.04 | 0.9643 | 0.9513 | 0.9436 |