Semantic Anomaly Detection on Osterhout and Mobley Semantic
100AccuracyMasked language model
Evaluation results
| Method | Links | |
|---|---|---|
| Masked language modelBackbone=BERT, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 100 | |
| Masked language modelBackbone=RoBERTa, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 98.1 | |
| Gaussian anomaly modelBackbone=RoBERTa, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 90.6 | |
| Masked language modelBackbone=XLNet, Anomaly Model=MLM, Layer Selection=Last layer2021.05 | 77.4 | |
| Gaussian anomaly modelBackbone=BERT, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 52.8 | |
| Gaussian anomaly modelBackbone=XLNet, Anomaly Model=GM, Layer Selection=Best layer2021.05 | 30.2 |