List your company

Misalignment Detection on Covert code sabotage Bloom behaviors (in-distribution)

0.959AUROC18-indicator ensemble
0.90180.916650.93150.94635Jun 23, 2026

Evaluation results

MethodLinks
18-indicator ensembleLayer=272026.060.959
18-indicator ensembleLayer=292026.060.956
RepE baselineLayer=272026.060.943
RepE baselineLayer=292026.060.904