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

Safety Evaluation on XSTest (Accuracy and Intervention Rate)

96Accuracy

Guidance Model

-3.8422.084873.92Apr 22, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
96-
2026.04
96-
2026.04
94-
2026.04
94-
2026.04
60-
2026.04
60-
2026.04
15-
2026.04
15-
2026.04
1520.3
2026.04
15-
2026.04
14-
2026.04
14-
2026.04
14-
2026.04
1428
2026.04
1222.2
2026.04
1222.2
2026.04
1020.3
2026.04
1028
2026.04
827.5
2026.04
827.5
2026.04
433.1
2026.04
333.1
2026.04
234.3
2026.04
234.3
2026.04
1-
2026.04
1-
2026.04
1-
2026.04
0-
2026.04
0-
2026.04
0-