Natural Language Inference on MultiNLI
82.4AccuracyERM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ERM2024.11 | 82.4 | 67.9 | — | |
| DFR2025.11 | 81.7 | 68.5 | — | |
| Mixup2025.11 | 81.4 | 68.5 | — | |
| GroupDRO2025.11 | 81.1 | 76 | — | |
| ReWeight2025.11 | 81 | 68.8 | — | |
| ERM2025.11 | 80.9 | 66.8 | — | |
| JTT2025.11 | 80.9 | 69.1 | — | |
| BSoftmax2025.11 | 80.9 | 66.9 | — | |
| SqrtReWeight2025.11 | 80.7 | 69.5 | — | |
| Focal2025.11 | 80.7 | 69.4 | — | |
| LDAM2025.11 | 80.7 | 69.6 | — | |
| CBLoss2025.11 | 80.6 | 72.2 | — | |
| SCER2025.11 | 80.4 | 76.8 | — | |
| LISA2025.11 | 80.3 | 73.3 | — | |
| JTT2024.11 | 80 | 70.2 | — | |
| Bias-Conditioned Self-Influencebase_method=JTT2024.11 | 79.8 | 73.6 | — | |
| MMD2025.11 | 78.8 | 69.1 | — | |
| IRM2025.11 | 77.8 | 63.6 | — | |
| ReSample2025.11 | 77.2 | 72.3 | — | |
| CVaRDRO2025.11 | 75.1 | 63 | — | |
| LfF2025.11 | 71.7 | 63.6 | — | |
| PDE2025.11 | 69.1 | 65.8 | — | |
| ElRep2025.11 | 69 | 66.8 | — | |
| BERTScoreThreshold=fixed midpoint threshold2026.05 | — | — | 33.33 | |
| BLEURTThreshold=fixed midpoint threshold2026.05 | — | — | 54.71 | |
| EmbSimThreshold=fixed midpoint threshold2026.05 | — | — | 33.33 | |
| MATCHAThreshold=fixed midpoint threshold2026.05 | — | — | 74.51 | |
| METEORThreshold=fixed midpoint threshold2026.05 | — | — | 50.51 | |
| R1-F1Threshold=fixed midpoint threshold2026.05 | — | — | 58.15 | |
| R2-F1Threshold=fixed midpoint threshold2026.05 | — | — | 43.15 | |
| RL-F1Threshold=fixed midpoint threshold2026.05 | — | — | 50.9 | |
| SimCSEThreshold=fixed midpoint threshold2026.05 | — | — | 33.71 |