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

Refusal on Refusal benchmark

97.81Alignment Probability

INNSTEER

32.383649.369366.35583.3407Jun 7, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
97.815.54
2026.06
91.93.63
2026.06
86.462.92
2026.06
84.65.2
2026.06
82.45.45
2026.06
82.15.35
2026.06
81.55.3
2026.06
81.093.65
2026.06
80.755.1
2026.06
78.475.24
2026.06
76.64.9
2026.06
76.524.89
2026.06
75.254.93
2026.06
75.14.95
2026.06
73.53.27
2026.06
72.33.7
2026.06
71.23.32
2026.06
70.43.3
2026.06
69.93.29
2026.06
69.53.42
2026.06
69.13.28
2026.06
68.33.1
2026.06
67.83.09
2026.06
66.93.08
2026.06
66.12.86
2026.06
65.122.92
2026.06
65.113.16
2026.06
64.83.22
2026.06
64.683.21
2026.06
63.353.24
2026.06
63.23.25
2026.06
63.22.89
2026.06
62.52.88
2026.06
61.82.87
2026.06
61.452.96
2026.06
60.92.86
2026.06
59.63.02
2026.06
59.183
2026.06
58.33.04
2026.06
58.13.05
2026.06
56.22.81
2026.06
53.52.83
2026.06
53.112.82
2026.06
522.84
2026.06
51.82.85
2026.06
48.83.18
2026.06
46.13.22
2026.06
45.23.2
2026.06
44.73.19
2026.06
43.83.18
2026.06
38.583.11
2026.06
36.23.12
2026.06
35.783.11
2026.06
35.053.14
2026.06
34.93.15