Jailbreak Attack on AdvBench (Avg.Q)
93.48Avg.QTAP
Evaluation Results
| Method | Links | |
|---|---|---|
| TAPTarget LLM=Claude-3.72026.02 | 93.48 | |
| TAPTarget LLM=Gemini-2.5-flash2026.02 | 93.14 | |
| TAPTarget LLM=Qwen3-235b2026.02 | 90.42 | |
| TAPTarget LLM=Deepseek-Reasoner2026.02 | 86.44 | |
| GPTFUZZERTarget LLM=GPT-4o2026.02 | 77.26 | |
| TAPTarget LLM=GPT-4o2026.02 | 65.72 | |
| PAIRTarget LLM=Gemini-2.5-flash2026.02 | 60 | |
| PAIRTarget LLM=GPT-4o2026.02 | 57.36 | |
| PAIRTarget LLM=Qwen3-235b2026.02 | 57 | |
| GPTFUZZERTarget LLM=Gemini-2.5-flash2026.02 | 56.96 | |
| TAPTarget LLM=Grok-32026.02 | 53.96 | |
| PAIRTarget LLM=Grok-32026.02 | 51.36 | |
| PAIRTarget LLM=Claude-3.72026.02 | 51.12 | |
| PAIRTarget LLM=Deepseek-Reasoner2026.02 | 40.32 | |
| GPTFUZZERTarget LLM=Claude-3.72026.02 | 32.62 | |
| CL-GSOTarget LLM=Claude-3.72026.02 | 21.42 | |
| GPTFUZZERTarget LLM=Qwen3-235b2026.02 | 19.08 | |
| AutoDAN-turbo-RTarget LLM=GPT-4o2026.02 | 16.84 | |
| AutoDAN-turbo-RTarget LLM=Claude-3.72026.02 | 14.8 | |
| AutoDAN-turbo-RTarget LLM=Grok-32026.02 | 13.58 | |
| AutoDAN-turbo-RTarget LLM=Qwen3-235b2026.02 | 13.48 | |
| AutoDAN-turbo-RTarget LLM=Deepseek-Reasoner2026.02 | 10.62 | |
| AutoDAN-turbo-RTarget LLM=Gemini-2.5-flash2026.02 | 10 | |
| GPTFUZZERTarget LLM=Deepseek-Reasoner2026.02 | 5.98 | |
| CL-GSOTarget LLM=Qwen3-235b2026.02 | 5.06 | |
| CL-GSOTarget LLM=GPT-4o2026.02 | 4 | |
| CL-GSOTarget LLM=Gemini-2.5-flash2026.02 | 3.62 | |
| CL-GSOTarget LLM=Deepseek-Reasoner2026.02 | 3.26 | |
| CC-BOSTarget LLM=Claude-3.72026.02 | 2.38 | |
| CC-BOSTarget LLM=Qwen3-235b2026.02 | 1.54 | |
| CC-BOSTarget LLM=Gemini-2.5-flash2026.02 | 1.46 | |
| GPTFUZZERTarget LLM=Grok-32026.02 | 1.32 | |
| CC-BOSTarget LLM=GPT-4o2026.02 | 1.28 | |
| CL-GSOTarget LLM=Grok-32026.02 | 1.24 | |
| CC-BOSTarget LLM=Grok-32026.02 | 1.18 | |
| CC-BOSTarget LLM=Deepseek-Reasoner2026.02 | 1.12 |