List your company

Safety & Helpfulness Evaluation on XSTest

74.8XSTest ScoreDefault
37.3647.0856.866.52Feb 16, 2026

Evaluation results

MethodLinks
DefaultModel=Qwen2.5-14B2026.0274.8
CBModel=Llama3-8B2026.0267.2
DefaultModel=Llama3-8B2026.0266
CATModel=Qwen2.5-14B2026.0254.4
Diffusion-onlyModel=Llama3-8B2026.0253.6
CATModel=Llama3-8B2026.0246.4
DATModel=Llama3-8B2026.0246.4
DATModel=Qwen2.5-14B2026.0246.4
MixAT-GCGModel=Llama3-8B2026.0245.6
LATModel=Llama3-8B2026.0244.4
MixAT-GCGModel=Qwen2.5-14B2026.0238.8