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

Behavioral Steering on Anthropic Model-Written Evaluations Sycophancy (test)

70Average Token Probability

CAA

21.1233.8146.559.19Jun 13, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
70
2026.06
65
2026.06
62
2026.06
56
2026.06
47
2026.06
32
2026.06
23