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

Strategic Deception

Benchmarks

Task NameDataset NameSOTA ResultTrend
Deception EvaluationStrategic Deception
CoT Plan Fidelity28.86
6
Misalignment DetectionStrategic deception Bloom behaviors (in-distribution)
AUROC0.817
4
Showing 2 of 2 rows