Jailbreaking on AdvBench (test)
72ASR (GPT-3.5)
MTSA-R3
Updated 4d ago
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
2025.05 | 72 | 66.5 | 56 | 50.5 | 64 | 74.5 | 63.92 | |
2025.05 | 68.5 | 64.5 | 56.5 | 47.5 | 58.5 | 71 | 61.1 | |
2025.05 | 64.5 | 65 | 54.5 | 46.5 | 51 | 67 | 58.08 | |
2025.05 | 63 | 58.5 | 53 | 43.5 | 45 | 57.5 | 53.42 | |
2025.05 | 57.5 | 61 | 51.5 | 20.5 | 39.5 | 61 | 48.5 | |
2025.05 | 52 | 63.5 | 55 | 24.5 | 48 | 63 | 51 | |
2025.05 | 47.5 | 51.5 | 44.5 | 35 | 41.5 | 43.5 | 43.92 | |
2025.05 | 44.5 | 56 | 38.5 | 37 | 43.5 | 43 | 43.75 | |
2025.05 | 41.5 | 32 | 28.5 | 22 | 38 | 32 | 32.33 | |
2025.05 | 36 | 24.5 | 14.5 | 26 | 32.5 | 28 | 26.91 | |
2025.05 | 34.5 | 39 | 33 | 16.5 | 21 | 16 | 26.68 | |
2025.05 | 33.5 | 12.5 | 22 | 34.5 | 24.5 | 36 | 27.17 |