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

Dishonesty Evaluation on Insecure code (test)

48.91Benchmark Dishonesty

PRISM

48.22452.854557.48562.1155May 20, 2026
Updated 13d ago

Evaluation Results

MethodLinks
2026.05
48.91
2026.05
49.27
2026.05
49.27
2026.05
50.73
2026.05
51.1
2026.05
51.1
2026.05
51.83
2026.05
52.56
2026.05
54.02
2026.05
57.66
2026.05
57.66
2026.05
58.76
2026.05
58.76
2026.05
59.12
2026.05
59.12
2026.05
59.85
2026.05
60.22
2026.05
60.95
2026.05
60.95
2026.05
61.68
2026.05
61.68
2026.05
62.41
2026.05
62.41
2026.05
62.41
2026.05
62.77
2026.05
62.77
2026.05
63.14
2026.05
63.14
2026.05
64.23
2026.05
64.96
2026.05
64.96
2026.05
66.06