List your company
ResearchFormerly WizWand

Safety Behavior Prediction on Assistant Turns

88.7Production RateDeployment Simulation
-3.44420.47844.468.322Jul 8, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Sexual content2026.0788.71.2100.3——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Harassment2026.0770.41.334.2——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Sexual content2026.0769.41.358.4——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Harassment2026.07561.137.6——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Sexual content2026.07531.756.2——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Hate2026.07382.137.9——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Violent wrongdoing2026.0735.21.826.1——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Harassment2026.07322.227.5——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Non-violent wrongdoing2026.07311.327.9——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Violent wrongdoing2026.0724.91.626.2——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Non-violent wrongdoing2026.0723.61.830.4——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Violent wrongdoing2026.0719.71.314.8——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Hate2026.0717.82.214.3——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Non-violent wrongdoing2026.0713.21.219.7——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Hate2026.0711.43.312.2——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Violence2026.0711.41.29.5——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Violence2026.0711.3110——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Sexual minors2026.079.41.312.6——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Violence2026.077.71.55.8——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Sexual minors2026.077.21.77.3——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Self-harm2026.075.615.8——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Self-harm2026.075.51.17.3——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Self-harm2026.0751.46.2——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Sexual minors2026.072.43.93——
Deployment SimulationDeployment=GPT-5.1 → GPT-5.2, Category=Extremism2026.0722.51.9——
Deployment SimulationDeployment=GPT-5.2 → GPT-5.4, Category=Extremism2026.070.85.60.8——
Deployment SimulationDeployment=GPT-5 → GPT-5.1, Category=Extremism2026.070.117.90.7——