List your company

Safety Evaluation on Anchor Safety Dataset

100Anchor AccuracyORIGINAL
-4235077Mar 19, 2026

Evaluation results

MethodLinks
ORIGINALModel Family=GEMMA-2-2B2026.03100
CDAModel Family=GEMMA-2-2B2026.03100
UGIDModel Family=GEMMA-2-2B2026.03100
ORIGINALModel Family=QWEN2.5-3B2026.03100
CDAModel Family=QWEN2.5-3B2026.03100
UGIDModel Family=QWEN2.5-3B2026.03100
ORIGINALModel Family=QWEN2.5-7B2026.03100
CDAModel Family=QWEN2.5-7B2026.03100
UGIDModel Family=QWEN2.5-7B2026.03100
ORIGINALModel Family=QWEN2.5-14B2026.03100
KLAADModel Family=QWEN2.5-14B2026.03100
UGIDModel Family=QWEN2.5-14B2026.03100
KLAADModel Family=GEMMA-2-2B2026.0350
KLAADModel Family=QWEN2.5-3B2026.0350
CDAModel Family=QWEN2.5-14B2026.0350
KLAADModel Family=QWEN2.5-7B2026.030