Hate Speech Detection on Twitter Offensive
84.5AccuracyFull dataset tuning
Evaluation results
| Method | Links | |
|---|---|---|
| Full dataset tuningTraining=Full dataset (FD)2023.05 | 84.5 | |
| Training-free Weight Recycling (L^T_KD)Distillation Data=Unlabeled Task Data, Source Dataset=Hate Speech2023.05 | 77.1 | |
| Training-free Weight Recycling (L^W_KD)Distillation Data=Wikipedia, Source Dataset=Hate Speech2023.05 | 74.5 | |
| 32-shot tuningTraining=32-shot (FS)2023.05 | 71.4 | |
| Demonstration learningMode=Zero-shot with context2023.05 | 63.5 |