Token-Level Anomaly Detection on Grammar (test)
63.71AUROCTokenCore
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TokenCoreBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 63.71 | 4.07 | |
| AEBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 61.9 | 3.86 | |
| LUNARBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 60.53 | 3.76 | |
| LOFBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 56.41 | 3.39 | |
| iForestBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 50.24 | 3.03 | |
| D.SVDDBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 47.98 | 3.12 | |
| ECODBackbone=BERT-base-uncased, Aggregation Strategy=max pooling, Evaluation Level=Token2026.01 | 44.57 | 2.68 |