Vulnerability Detection on REVEAL
89.9AccuracyWhite-Basilisk
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| White-BasiliskTraining Strategy=Dataset-Specific2025.07 | 89.9 | 49.3 | — | — | — | — | — | — | |
| White-BasiliskTraining Strategy=Unified2025.07 | 88.8 | 47 | — | — | — | — | — | — | |
| VulStyle2026.04 | 85.22 | 45.27 | 139 | 1,799 | 240 | 96 | 36.67 | 59.14 | |
| VulBERTa-MLP2025.07 | 84.5 | 45.3 | — | — | — | — | — | — | |
| VulBERTa-MLP2026.04 | 84.48 | 45.27 | 146 | 1,775 | 269 | 84 | 35.18 | 63.48 | |
| REVEAL2025.07 | 84.4 | 41.3 | — | — | — | — | — | — | |
| CodeBERTvariant=Baseline2026.04 | 83.46 | 42.33 | 138 | 1,760 | 279 | 97 | 33.09 | 58.72 | |
| UniXcodervariant=Baseline2026.04 | 81.31 | 42.79 | 159 | 1,690 | 349 | 76 | 31.29 | 67.65 | |
| VulBERTa-CNN2026.04 | 79.73 | 42.59 | 171 | 1,642 | 402 | 59 | 29.84 | 74.35 | |
| VulBERTa-CNN2025.07 | 79.7 | 42.6 | — | — | — | — | — | — | |
| Baseline-BiLSTM2025.07 | 77.1 | 39.1 | — | — | — | — | — | — | |
| Baseline-TextCNN2025.07 | 73.2 | 37.4 | — | — | — | — | — | — |