Jailbreak Attack on Crescendo (test)
37ASRClaude Sonnet 4.5
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude Sonnet 4.5Safety Steering=Disabled2025.02 | 37 | |
| GPT-5Safety Steering=Disabled2025.02 | 35 | |
| Claude Sonnet 4.5Safety Steering=Enabled2025.02 | 30 | |
| GPT-5Safety Steering=Enabled2025.02 | 15 |