Passivization on Passivization Hierarchical Generalization
100Partial AccuracyT5
Evaluation results
| Method | Links | ||
|---|---|---|---|
| T5Large pre-trained model=true2025.11 | 100 | — | |
| mT5Large pre-trained model=true2025.11 | 100 | — | |
| BARTLarge pre-trained model=true2025.11 | 100 | — | |
| TfTraining=Language Modeling Objective2025.11 | 100 | — | |
| mBARTLarge pre-trained model=true, Overtrained=true2025.11 | 80 | — | |
| Tf+Sup+SupArchitecture=Transformer, Supervision=Double2025.11 | 75.8 | 0 | |
| Tf+NdArchitecture=Stack Attention Transformer2025.11 | 61.2 | 0 | |
| Tf+Nd+NdArchitecture=Stack Attention Transformer, Stack=Double2025.11 | 56.9 | 0 | |
| Tf+SupArchitecture=Transformer, Supervision=Yes2025.11 | 56.5 | 0 | |
| TfArchitecture=Transformer2025.11 | 52.6 | 0 | |
| LSTM+Nd+RArchitecture=Stack LSTM, Variant=R2025.11 | 18.4 | 0 | |
| LSTM+SupArchitecture=LSTM, Supervision=Yes2025.11 | 8.9 | 0 | |
| LSTM+Sup+RArchitecture=LSTM, Supervision=Yes, Variant=R2025.11 | 7.1 | 0 | |
| RNNArchitecture=RNN2025.11 | 6.1 | 0 | |
| RNN+Nd+RArchitecture=Stack RNN, Variant=R2025.11 | 5.7 | 0 | |
| LSTMArchitecture=LSTM2025.11 | 5.7 | 0 | |
| RNN+SupArchitecture=RNN, Supervision=Yes2025.11 | 5.6 | 0 | |
| LSTM+NdArchitecture=Stack LSTM2025.11 | 5.6 | 0 | |
| RNN+Sup+RArchitecture=RNN, Supervision=Yes, Variant=R2025.11 | 4.8 | 0 | |
| RNN+NdArchitecture=Stack RNN2025.11 | 4.7 | 0 |