Tense reinflection on Tense reinflection (auxiliary do) 1.0 (Linear generalization)
87.1Full AccuracyTf+Sup+Sup
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Tf+Sup+SupArchitecture=Transformer, Augmentation=Dual Supervision2025.11 | 87.1 | 87.1 | |
| TfArchitecture=Transformer2025.11 | 85.2 | 85.4 | |
| Tf+Nd+NdArchitecture=Transformer, Augmentation=Dual Neural Data Structure2025.11 | 82.9 | 83.1 | |
| Tf+SupArchitecture=Transformer, Augmentation=Supervision2025.11 | 62.9 | 63 | |
| Tf+NdArchitecture=Transformer, Augmentation=Neural Data Structure2025.11 | 52.1 | 52.3 | |
| RNNArchitecture=RNN2025.11 | 43.2 | 91.8 | |
| RNN+Sup+RArchitecture=RNN, Augmentation=Supervision + R2025.11 | 17.7 | 89.1 | |
| RNN+NdArchitecture=RNN, Augmentation=Neural Data Structure2025.11 | 14.3 | 93.4 | |
| RNN+SupArchitecture=RNN, Augmentation=Supervision2025.11 | 8 | 91.2 | |
| RNN+Nd+RArchitecture=RNN, Augmentation=Neural Data Structure + R2025.11 | 7 | 90.5 | |
| LSTM+Nd+RArchitecture=LSTM, Augmentation=Neural Data Structure + R2025.11 | 2.2 | 86.8 | |
| LSTM+SupArchitecture=LSTM, Augmentation=Supervision2025.11 | 0.3 | 43.1 | |
| LSTMArchitecture=LSTM2025.11 | 0.2 | 46.3 | |
| LSTM+Sup+RArchitecture=LSTM, Augmentation=Supervision + R2025.11 | 0.2 | 45.9 | |
| LSTM+NdArchitecture=LSTM, Augmentation=Neural Data Structure2025.11 | 0.1 | 49.2 |