Toxicity Detection
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
51.43Recall
4
May 26, 2026
0.845Overall ROC AUC
4
Mar 18, 2026
61Precision
4
Feb 26, 2026
63Precision
4
Feb 26, 2026
0.977Accuracy
4
Feb 26, 2026
0.967Accuracy
4
Feb 26, 2026
88.15Accuracy
4
Feb 26, 2026
100Bypass Rate
3
May 25, 2026
100Accuracy
3
Feb 26, 2026
2.71Error Rate
3
Feb 26, 2026
93.85Detection Rate
3
Feb 26, 2026
93.64Detection Rate
3
Feb 26, 2026
0.973Detection Rate
3
Feb 26, 2026
80.06Accuracy
1
Mar 17, 2026
69.502Accuracy
1
Mar 17, 2026
85.81Accuracy
1
Mar 17, 2026
—
—Primary metric
0
Feb 26, 2026