Commonsense Anomaly Detection on Urbach and Kutas Commonsense
93.9AccuracyMasked language model
Evaluation results
| Method | Links | |
|---|---|---|
| Masked language modelBackbone=RoBERTa, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 93.9 | |
| Masked language modelBackbone=BERT, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 92.4 | |
| Masked language modelBackbone=XLNet, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 71.2 | |
| Gaussian anomaly modelBackbone=XLNet, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 50 | |
| Gaussian anomaly modelBackbone=RoBERTa, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 48.5 | |
| Gaussian anomaly modelBackbone=BERT, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 47 |