Safety Classification
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
100Accuracy
7
Mar 5, 2026
77.65Precision
7
Feb 26, 2026
93.9Accepted HF Accuracy
6
Jul 7, 2026
100Adv-suffix Accuracy
6
Jul 7, 2026
99MCA
6
Apr 21, 2026
99MCA Accuracy
6
Apr 21, 2026
100MCA
6
Apr 21, 2026
100Acc MCA
6
Apr 21, 2026
98.2AUROC
6
Feb 26, 2026
86.2XSTest Accuracy
5
Jun 16, 2026
84.4XSTest Accuracy
5
Jun 16, 2026
82.9XSTest Accuracy
5
Jun 16, 2026
88.8XSTest Accuracy
5
Jun 16, 2026
85.82AUROC
5
May 26, 2026
0.718AUROC
5
May 26, 2026
98.2Accuracy
5
Apr 21, 2026
85.5F1 Score
4
Jun 11, 2026
94.7Animal Safety Classification Accuracy
4
May 12, 2026
0.107Inference Speed (s/sample)
4
May 12, 2026
97.75F1 (off)
4
Mar 23, 2026
98.38F1 (off)
4
Mar 23, 2026
0.56Violation Rate
3
May 15, 2026
0.19Violation Rate
3
May 15, 2026
0.07Violation Rate
3
May 15, 2026
85.6Accuracy
2
May 26, 2026