Red-teaming Safety Evaluation on Harmbench
0.3ASR
SafeKey
Updated 4d ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.01 | 0.3 | - | - | - | |
2026.01 | 0.5 | - | - | - | |
2026.01 | 0.8 | - | - | - | |
| 1 | 2 | 1.64 | - | ||
2026.01 | 1 | - | - | - | |
2026.01 | 1.3 | - | - | - | |
2026.01 | 1.5 | - | - | - | |
2026.01 | 1.5 | - | - | - | |
| 2 | 12 | 1.65 | - | ||
| 2 | 6 | 1.64 | - | ||
2026.01 | 3 | - | - | - | |
2026.01 | 3.5 | - | - | - | |
| 4 | 15 | 1.87 | - | ||
2026.01 | 4.8 | - | - | - | |
2026.01 | 5 | - | - | - | |
2026.01 | 6.8 | - | - | - | |
| 7 | 46 | 1.91 | - | ||
| 7 | 27 | 2.04 | - | ||
| 9 | 36 | 2.37 | - | ||
2026.01 | 9 | - | - | - | |
2026.01 | 10.5 | - | - | - | |
2026.01 | 12 | - | - | - | |
| 19 | 41 | 2.29 | - | ||
2026.01 | 30.8 | - | - | - | |
2026.01 | 35.5 | - | - | - | |
2026.01 | 36.3 | - | - | - | |
2026.01 | 38 | - | - | - | |
2026.01 | 39.5 | - | - | - | |
2026.01 | 40.8 | - | - | - | |
2026.01 | 49.3 | - | - | - | |
2026.01 | 51.5 | - | - | - | |
2026.01 | 53.3 | - | - | - |