List your company

Safety Evaluation on ActorAttack

3.5ASRGPT-4o
1.2416.49531.7547.005Feb 28, 2025

Evaluation results

MethodLinks
GPT-4oNBF steering=true2025.023.5
GPT-3.5-turboNBF steering=true2025.024
Claude 3.5 SonnetNBF steering=true2025.024
o1NBF steering=true2025.029
Claude 3.5 SonnetNBF steering=false2025.0220
o1NBF steering=false2025.0251
GPT-3.5-turboNBF steering=false2025.0258.5
GPT-4oNBF steering=false2025.0260