Commonsense Anomaly Detection on Warren Commonsense
75AccuracyGaussian anomaly model
Evaluation results
| Method | Links | |
|---|---|---|
| Gaussian anomaly modelBackbone=RoBERTa, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 75 | |
| Gaussian anomaly modelBackbone=BERT, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 60 | |
| Masked language modelBackbone=XLNet, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 60 | |
| Masked language modelBackbone=BERT, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 55 | |
| Masked language modelBackbone=RoBERTa, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 45 | |
| Gaussian anomaly modelBackbone=XLNet, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 30 |