List your company

Structural Erasure on WMDP chem

1.351CAD ScoreSimNPO
-0.054040.310730.67551.04027May 14, 2026

Evaluation results

MethodLinks
SimNPOModel=Llama-3.1-8B-Instruct2026.051.3511.3510.875
GU+SimNPOModel=Llama-3.1-8B-Instruct2026.051.2921.2920.833
MANSUModel=Llama-3.1-8B-Instruct2026.051.0970.3980.141
Global GAModel=Qwen-3-8B2026.051.0571.0570.789
MANSUModel=Qwen-3-8B2026.051.0410.4870.198
SimNPOModel=Qwen-3-8B2026.050.9660.9660.724
NPOModel=Llama-3.1-8B-Instruct2026.050.8360.8360.57
GU+SimNPOModel=Qwen-3-8B2026.050.7930.7930.543
NPOModel=Qwen-3-8B2026.050.7730.7730.551
Surgical GAModel=Qwen-3-8B2026.050.6120.6120.325
Global GAModel=Llama-3.1-8B-Instruct2026.050.40.40.282
Surgical GAModel=Llama-3.1-8B-Instruct2026.050.3560.3560.172
LUNARModel=Llama-3.1-8B-Instruct2026.050.0330.9740.256
LUNARModel=Qwen-3-8B2026.050.0310.9370.247
Zero-shotModel=Llama-3.1-8B-Instruct2026.05000
Zero-shotModel=Qwen-3-8B2026.05000