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

Safety and Helpfulness Evaluation on VLGuard

95.1Safety Score

GRPO

34.57250.2866681.714Mar 18, 2026Mar 30, 2026Apr 12, 2026Apr 25, 2026May 7, 2026May 20, 2026Jun 2, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
95.1-51.79
2026.03
88.4881.82-
2026.06
88.2-77.6
2026.06
87.9-55.38
2026.06
86.3-97.13
2026.03
84.277.6-
2026.06
83.6-94.62
2026.06
83-96.42
2026.03
82.674.51-
2026.06
81.9-95.34
2026.06
81.8-95.52
2026.06
81.1-94.8
2026.03
80.176.7-
2026.03
79.8877.88-
2026.03
79.6273.25-
2026.03
79.5978.41-
2026.03
77.1168.67-
2026.03
76.966.6-
2026.03
75.175.3-
2026.03
74.371-
2026.06
74.3-94
2026.06
73.7-95.52
2026.03
71.972.37-
2026.03
71.5466.13-
2026.06
70.4-94.98
2026.06
69.6-64.87
2026.06
67.6-97.31
2026.03
66.256.3-
2026.03
59.3463.72-
2026.03
52.846.2-
2026.03
48.239-
2026.03
36.937.8-