Jailbreak Attack on Harmbench Target Gemini-2.0-flash
78.33Attack Success Rate (ASR)Babel
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Babel2026.05 | 78.33 | 38 | |
| PiF2026.05 | 72.67 | 25 | |
| FlipAttack2026.05 | 60 | 1 | |
| TAP2026.05 | 38.67 | 41 | |
| Bonquery count limit=502026.05 | 15.67 | 43 |
| Method | Links | ||
|---|---|---|---|
| Babel2026.05 | 78.33 | 38 | |
| PiF2026.05 | 72.67 | 25 | |
| FlipAttack2026.05 | 60 | 1 | |
| TAP2026.05 | 38.67 | 41 | |
| Bonquery count limit=502026.05 | 15.67 | 43 |