Toxicity Detection on L2DTnH (test)
96.05AccuracyIGC-BERT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| IGC-BERTMode=fine-tuned2026.04 | 96.05 | 8,571 | 68.57 | 76.19 | |
| ChatGPT 4oVersion=June 2025, Prompt=P22026.04 | 91.78 | 0.3662 | 66.67 | 47.27 | |
| unitary_toxic-bertVariant=base2026.04 | 91.66 | 0.5401 | 65 | 59 | |
| garak_llm_roberta_toxicity_classifier2026.04 | 90.74 | 0.4978 | 39.64 | 44.14 | |
| martin_ha_toxic_comment_model2026.04 | 90.68 | 0.4908 | 28.57 | 36.12 | |
| protectai_unbiased_toxic_roberta_onnx2026.04 | 90.21 | 0.4659 | 41.43 | 43.86 | |
| ChatGPT 4oVersion=June 2025, Prompt=P12026.04 | 90.08 | 0.3495 | 92.31 | 50.7 | |
| junglelee_bert_toxic_comment_classification2026.04 | 87.22 | 0.3977 | 75 | 51.98 | |
| nicholas_kluge_toxicity_model2026.04 | 87.08 | 0.3866 | 68.21 | 49.35 | |
| Llama 3.22026.04 | 75.42 | 0.2451 | 80 | 37.52 |