Question Answering on AddOneSent (test)
64.4EMmulti-domain debiasing framework
Evaluation results
| Method | Links | ||
|---|---|---|---|
| multi-domain debiasing frameworkSetting=Five Domains, Backbone=ELECTRA-small2026.01 | 64.4 | 71.5 | |
| ELECTRA-small (Baseline)Setting=Baseline, Backbone=ELECTRA-small2026.01 | 62.5 | 69.1 | |
| FusionNetmodel_type=Single, use_CoVe=false2017.11 | 55.2 | 61.2 | |
| FusionNetmodel_type=Single, selection_criteria=10-run best2017.11 | 54.8 | 60.9 | |
| FusionNetmodel_type=Ensemble2017.11 | 54.7 | 60.7 | |
| FusionNetModel Type=Ensemble2017.11 | 54.7 | 60.7 | |
| FusionNetmodel_type=Single, selection_criteria=10-run mean2017.11 | 53.1 | 59.3 | |
| Previous SotAmodel_type=Ensemble2017.11 | 48.7 | 55.3 | |
| Mnemonic ReaderModel Type=Ensemble2017.11 | 48.7 | 55.3 | |
| Mnemonic ReaderModel Type=Single2017.11 | 48.5 | 56 | |
| ReasoNetModel Type=Ensemble2017.11 | 43.6 | 49.8 | |
| BiDAFModel Type=Ensemble2017.11 | 40.7 | 46.9 | |
| SEDTModel Type=Ensemble2017.11 | 40 | 46.5 | |
| Match-LSTMModel Type=Ensemble2017.11 | 34.8 | 41.8 | |
| LR BaselineModel Type=Single2017.11 | 22.3 | 30.4 |