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

Refusal Evaluation on OR-Bench

99.6Toxic Refusal Rate

SFT (30/70)

74.43280.96687.594.034Sep 30, 2025
Updated 3d ago

Evaluation Results

MethodLinks
2025.09
99.691.8
2025.09
99.598.5
2025.09
99.389
2025.09
9988
2025.09
98.984.7
2025.09
98.794.9
2025.09
9870.5
2025.09
97.569.2
2025.09
97.273.6
2025.09
97.170.7
2025.09
96.966.2
2025.09
96.971.1
2025.09
96.770.5
2025.09
96.666.8
2025.09
96.466.8
2025.09
96.177.9
2025.09
94.857.9
2025.09
94.342.6
2025.09
94.150.8
2025.09
93.248.3
2025.09
92.848.8
2025.09
92.543.5
2025.09
92.545.1
2025.09
92.343.1
2025.09
9242.8
2025.09
91.980.3
2025.09
91.972.4
2025.09
91.443.2
2025.09
91.138.8
2025.09
90.233.2
2025.09
88.528.9
2025.09
87.929.7
2025.09
86.734.7
2025.09
85.184.8
2025.09
80.114.6
2025.09
79.813.1
2025.09
79.512.9
2025.09
79.315
2025.09
79.312.8
2025.09
75.412.2