List your company

Misalignment Detection on Sycophancy Bloom behaviors (in-distribution)

93.5AUROCRepE baseline
91.21291.80692.492.994Jun 23, 2026

Evaluation results

MethodLinks
RepE baselineLayer=292026.0693.5
18-indicator ensembleLayer=272026.0693
18-indicator ensembleLayer=292026.0692.8
RepE baselineLayer=272026.0691.3