List your company

Harmful Content Detection on ToxicChat

92.6Safe RateLlama Guard 3
33.00848.47963.9579.421Jul 16, 2025

Evaluation results

MethodLinks
Llama Guard 32025.0792.647.2
Latent Guard-Llama3Backbone=Llama32025.0783.531.7
Latent Guard-Qwen2Backbone=Qwen22025.0780.134
Latent Guard-Llama2Backbone=Llama22025.0735.372.7