List your company

Misalignment Detection on Strategic sandbagging Bloom behaviors (in-distribution)

0.985AUROC18-indicator ensemble
0.885160.911080.9370.96292Jun 23, 2026

Evaluation results

MethodLinks
18-indicator ensembleLayer=272026.060.985
18-indicator ensembleLayer=292026.060.98
RepE baselineLayer=272026.060.889
RepE baselineLayer=292026.060.889