Jailbreak Attack on Claude Sonnet API 3.5
80.5ASRTriageFuzz
Evaluation Results
| Method | Links | |
|---|---|---|
| TriageFuzzQuery Budget=252026.03 | 80.5 | |
| TriageFuzzQuery Budget=202026.03 | 72 | |
| TriageFuzzQuery Budget=152026.03 | 60.5 | |
| TriageFuzzQuery Budget=102026.03 | 47.5 | |
| PAIRQuery Budget=252026.03 | 28 | |
| TAPQuery Budget=252026.03 | 24.5 | |
| PAIRQuery Budget=202026.03 | 20.5 | |
| TAPQuery Budget=202026.03 | 17 | |
| PAIRQuery Budget=152026.03 | 13 | |
| TAPQuery Budget=152026.03 | 9.5 | |
| PAIRQuery Budget=102026.03 | 6.5 | |
| TAPQuery Budget=102026.03 | 3.5 | |
| GPTFuzzQuery Budget=252026.03 | 3 | |
| GPTFuzzQuery Budget=202026.03 | 1 | |
| GPTFuzzQuery Budget=102026.03 | 0 | |
| GPTFuzzQuery Budget=152026.03 | 0 |