Natural Language Inference on MultiNLI WILDS (test)
82.1IID AccuracyDFR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DFRBias labels used for training (TR.)=false, Bias labels used for validation (VAL)=true2025.12 | 82.1 | 74.7 | |
| ERMBias labels used for training (TR.)=false, Bias labels used for validation (VAL)=false2025.12 | 82.09 | 68.11 | |
| AFRBias labels used for training (TR.)=false, Bias labels used for validation (VAL)=false2025.12 | 81.4 | 73.4 | |
| GroupDROBias labels used for training (TR.)=true, Bias labels used for validation (VAL)=true2025.12 | 81.4 | 77.7 | |
| NSFBias labels used for training (TR.)=false, Bias labels used for validation (VAL)=false2025.12 | 80.43 | 73.12 | |
| JTTBias labels used for training (TR.)=false, Bias labels used for validation (VAL)=false2025.12 | 78.6 | 72.6 |