Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

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

0.985AUROC

18-indicator ensemble

0.885160.911080.9370.96292Jun 23, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
0.985
2026.06
0.98
2026.06
0.889
2026.06
0.889