Toxicity Classification on WZ
73.4Harmonic F1Back-Translate
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Back-TranslateClassifier=BERT2026.01 | 73.4 | 78.4 | |
| GPT4o-TCClassifier=RoBERTa2026.01 | 73.4 | 78.3 | |
| T5-ParaphaseClassifier=RoBERTa2026.01 | 73.1 | 78.3 | |
| Back-TranslateClassifier=RoBERTa2026.01 | 72.8 | 77.9 | |
| Llama3.2-TCClassifier=RoBERTa2026.01 | 72.8 | 77.7 | |
| ToxiGANClassifier=RoBERTa2026.01 | 72.8 | 77.9 | |
| GPT4o-TCClassifier=BERT2026.01 | 72.6 | 77.6 | |
| Ideal (Gold Label)Classifier=RoBERTa2026.01 | 72.5 | 77.9 | |
| GPT4.1-TCClassifier=RoBERTa2026.01 | 72.4 | 77.6 | |
| ToxiGANClassifier=BERT2026.01 | 72.2 | 77.7 | |
| Mistral-v0.3-SGClassifier=RoBERTa2026.01 | 72 | 77.5 | |
| EDAClassifier=BERT2026.01 | 71.9 | 77.2 | |
| Mistral-v0.3-SGClassifier=BERT2026.01 | 71.9 | 77.2 | |
| Llama3.2-TCClassifier=BERT2026.01 | 71.9 | 77.3 | |
| GPT4.1-TCClassifier=BERT2026.01 | 71.8 | 77.2 | |
| OversamplingClassifier=RoBERTa2026.01 | 71.7 | 77.3 | |
| Mistral-v0.3-FSClassifier=RoBERTa2026.01 | 71.7 | 77.2 | |
| Mistral-v0.3-FSClassifier=BERT2026.01 | 71.5 | 76.8 | |
| Mistral-v0.3-ZGClassifier=BERT2026.01 | 71.3 | 76.9 | |
| EDAClassifier=RoBERTa2026.01 | 71.3 | 77 | |
| Mistral-v0.3-ZGClassifier=RoBERTa2026.01 | 71.3 | 77 | |
| Base (No Aug.)Classifier=RoBERTa2026.01 | 71.2 | 76.9 | |
| T5-ParaphaseClassifier=BERT2026.01 | 70.4 | 76.3 | |
| Ideal (Gold Labels)Classifier=BERT2026.01 | 69.7 | 76 | |
| OversamplingClassifier=BERT2026.01 | 68.3 | 75 | |
| Base (No Aug.)Classifier=BERT2026.01 | 68.1 | 74.8 |