List your company

Bio-threat knowledge evaluation on WMDP Bio

72.43Accuracy (All)Llama 3.1 8B Instruct
52.004457.307262.6167.9128Aug 8, 2025

Evaluation results

MethodLinks
Llama 3.1 8B InstructParameters=8B, Instruct=true2025.0872.4389.1464.6346.28
Tulu-3 8BParameters=8B2025.0869.687.1661.4149.81
OLMO-2 7BParameters=7B, Instruct=false2025.0867.3286.6758.2942.94
OLMO-2 7B InstructParameters=7B, Instruct=true2025.0866.3887.6556.4545.35
Gemma-3 4B InstructParameters=4B, Instruct=true2025.0863.8685.6853.6938.1
Gemma-3 4BParameters=4B, Instruct=false2025.0861.0479.5152.4236.15
Our Unfiltered Model (Baseline)status=baseline2025.0852.7973.3343.236.34