List your company
ResearchFormerly WizWand

Safety Classification Disagreement on WildChat Content

0.4Disagreement Rate (per 1k Conversations)Gemini 3.1
-2.23215.53433.351.066May 22, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Gemini 3.1Category=Non-Violent Crime, Evaluation Protocol=Constitution2026.050.4
GPT-5.4 MiniCategory=Non-Violent Crime, Evaluation Protocol=Constitution2026.050.7
Gemini 3.1Category=Harassment, Evaluation Protocol=Constitution2026.051
GPT-5.4 MiniCategory=Harassment, Evaluation Protocol=Constitution2026.051.1
GPT-5.4Category=Harassment, Evaluation Protocol=Constitution2026.051.5
GPT-5.4Category=Non-Violent Crime, Evaluation Protocol=Constitution2026.051.5
Gemini 3.1Category=Hate Speech, Evaluation Protocol=Constitution2026.051.7
GPT-5.4Category=Hate Speech, Evaluation Protocol=Constitution2026.051.9
Safeguard 20BCategory=Harassment, Evaluation Protocol=Constitution2026.051.9
GPT-5.4 MiniCategory=Hate Speech, Evaluation Protocol=Constitution2026.052.2
Safeguard 20BCategory=Hate Speech, Evaluation Protocol=Constitution2026.052.4
Gemini 3.1Category=Hate Speech, Evaluation Protocol=Definition2026.052.5
Safeguard 20BCategory=Non-Violent Crime, Evaluation Protocol=Constitution2026.052.9
GPT-5.4 MiniCategory=Hate Speech, Evaluation Protocol=Definition2026.054.1
Safeguard 20BCategory=Hate Speech, Evaluation Protocol=Definition2026.054.1
GPT-5.4 NanoCategory=Hate Speech, Evaluation Protocol=Constitution2026.054.6
Gemini 3.1Category=Non-Violent Crime, Evaluation Protocol=Definition2026.054.8
Gemini 3.1Category=Harassment, Evaluation Protocol=Definition2026.055.1
GPT-5.4 NanoCategory=Non-Violent Crime, Evaluation Protocol=Constitution2026.055.1
GPT-5.4 MiniCategory=Non-Violent Crime, Evaluation Protocol=Definition2026.056.8
GPT-5.4 NanoCategory=Harassment, Evaluation Protocol=Constitution2026.056.8
GPT-5.4Category=Non-Violent Crime, Evaluation Protocol=Definition2026.058.1
GPT-5.4 NanoCategory=Hate Speech, Evaluation Protocol=Definition2026.058.1
Safeguard 20BCategory=Non-Violent Crime, Evaluation Protocol=Definition2026.059.7
GPT-5.4Category=Hate Speech, Evaluation Protocol=Definition2026.0512.1
GPT-5.4 NanoCategory=Non-Violent Crime, Evaluation Protocol=Definition2026.0520.5
Safeguard 20BCategory=Harassment, Evaluation Protocol=Definition2026.0529.3
GPT-5.4 MiniCategory=Harassment, Evaluation Protocol=Definition2026.0542.2
GPT-5.4Category=Harassment, Evaluation Protocol=Definition2026.0547.2
GPT-5.4 NanoCategory=Harassment, Evaluation Protocol=Definition2026.0566.2