Jailbreak Attack on TrotterStr
35Success Count @ k=1Mousetrap
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MousetrapTarget LRM=DeepSeek-R1, Strategy=TrotterStr2025.02 | 35 | 70 | 14 | 98 | 1 | 100 | 0 | 0 | |
| MousetrapTarget LRM=Gork-3, Strategy=TrotterStr2025.02 | 24 | 48 | 26 | 100 | 0 | 100 | 0 | 0 | |
| MousetrapTarget LRM=o3-mini, Strategy=TrotterStr2025.02 | 19 | 38 | 29 | 96 | 2 | 100 | 0 | 0 | |
| MousetrapTarget LRM=Gemini-2.0 (H), Strategy=TrotterStr2025.02 | 19 | 38 | 19 | 76 | 1 | 98 | 1 | 2 | |
| MousetrapTarget LRM=o1, Strategy=TrotterStr2025.02 | 16 | 32 | 31 | 94 | 3 | 100 | 0 | 0 | |
| MousetrapTarget LRM=o1-mini, Strategy=TrotterStr2025.02 | 15 | 30 | 21 | 72 | 12 | 96 | 2 | 4 | |
| MousetrapTarget LRM=QwQ-Plus, Strategy=TrotterStr2025.02 | 12 | 24 | 37 | 98 | 1 | 100 | 0 | 0 | |
| MousetrapTarget LRM=Gemini-2.5-Pro (H), Strategy=TrotterStr2025.02 | 9 | 18 | 40 | 98 | 1 | 100 | 0 | 0 | |
| MousetrapTarget LRM=Gemini-2.0 (M&H), Strategy=TrotterStr2025.02 | 8 | 16 | 21 | 58 | 6 | 70 | 15 | 30 | |
| MousetrapTarget LRM=Claude-3-7-Sonnet, Strategy=TrotterStr2025.02 | 4 | 8 | 43 | 94 | 3 | 100 | 0 | 0 | |
| MousetrapTarget LRM=Gemini-2.5-Pro (M&H), Strategy=TrotterStr2025.02 | 4 | 8 | 13 | 34 | 6 | 46 | 27 | 54 | |
| MousetrapTarget LRM=Claude-3-5-Sonnet, Strategy=TrotterStr2025.02 | 2 | 4 | 31 | 66 | 10 | 86 | 7 | 14 |