List your company

Safety Evaluation on Alpaca Mix Safe

0Refusal RateRandom feature
-1.146.55514.2521.945Jun 25, 2026

Evaluation results

MethodLinks
Random featureModel=Qwen3-1.7B-it2026.060
No SteeringModel=Qwen3-4B-it2026.060.8
No SteeringModel=Qwen3-1.7B-it2026.060.9
Random featureModel=Qwen3-4B-it2026.064.3
Target featureModel=Qwen3-1.7B-it2026.0610.6
Target featureModel=Qwen3-4B-it2026.0628.5