Jailbreaking on HarmBench (test)
97ASR (GPT-4o)
HMNS
Updated 4d ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2026.04 | 97 | 94 | 2 | |
2026.04 | 96 | 92 | 2.1 | |
2026.04 | 91 | 88 | 7.7 | |
2026.04 | 90 | 86 | 7.9 | |
2026.04 | 90 | 84 | 2.1 | |
2026.04 | 84 | 78 | 8.4 | |
2026.04 | 83.1 | 78.2 | 9.5 | |
2026.04 | 82 | 76.6 | 9.8 | |
2026.04 | 80 | 73.9 | 10 | |
2026.04 | 75.6 | 70.2 | 11.9 | |
2026.04 | 74.1 | 68.3 | 12.1 | |
2026.04 | 71 | 64.8 | 12.9 | |
2026.04 | 70.6 | 64.9 | 12.8 | |
2026.04 | 69.1 | 63.2 | 13.1 | |
2026.04 | 68.9 | 63.4 | 10.8 | |
2026.04 | 67.4 | 61.9 | 10.1 | |
2026.04 | 67 | 61.2 | 10.9 | |
2026.04 | 65.7 | 60.1 | 10.6 | |
2026.04 | 63.7 | 58.4 | 11 | |
2026.04 | 62.4 | 56.7 | 13.9 | |
2026.04 | 62.2 | 56.7 | 11.4 | |
2026.04 | 60.1 | 54.2 | 11.7 | |
2026.04 | 58.6 | 53.1 | 11.2 | |
2026.04 | 55.2 | 49.8 | 12.3 | |
2026.04 | 43.8 | 38.4 | 16.2 | |
2026.04 | 41.3 | 36.1 | 16.8 | |
2026.04 | 37.9 | 32.8 | 17.6 |