List your company

Jailbreak Attack on ActorAttack (test)

54ASRClaude Sonnet 4.5
215.52942.5Feb 28, 2025

Evaluation results

MethodLinks
Claude Sonnet 4.5Safety Steering=Disabled2025.0254
GPT-5Safety Steering=Disabled2025.0235.5
Claude Sonnet 4.5Safety Steering=Enabled2025.0211
GPT-5Safety Steering=Enabled2025.024