List your company

Flagging high human disagreement cases on Measuring Hate Speech corpus

40.1Precision (High Disagreement)Category-based escalation
33.02834.86436.738.536Apr 4, 2026

Evaluation results

MethodLinks
Category-based escalation2026.0440.184.554.8
Divergence Only2026.0434.791.550.3
Random Baseline2026.0433.350.540.1