Benchmark Quality Evaluation on VLSafetyBencher
6.73MADMLLMGuard
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MLLMGuard2026.01 | 6.73 | 34.32 | 46.38 | 82.64 | |
| DME+MLLMGuard2026.01 | 7.1 | 33.84 | 47.8 | 82.98 | |
| DataGen2026.01 | 7.18 | 29.11 | 49.88 | 77.66 | |
| AutoBencher2026.01 | 7.63 | 39.32 | 50.33 | 78.99 | |
| SafeBench2026.01 | 8.32 | 22.81 | 54.3 | 82.05 | |
| VLSafetyBencher2026.01 | 15.03 | 39.16 | 69.97 | 83.1 |