Personal Attack Detection on Personal Attack Detection
0.895Accuracyllama4-scout-17b
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| llama4-scout-17b2025.12 | 0.895 | 0.879 | 0.886 | 0.883 | 0.931 | |
| qwen3-32b2025.12 | 0.89 | 0.881 | 0.871 | 0.876 | 0.929 | |
| Our Modelarchitecture=GNN-based framework2025.12 | 0.887 | 0.862 | 0.889 | 0.875 | 0.957 | |
| qwen3-235b2025.12 | 0.885 | 0.824 | 0.945 | 0.88 | 0.942 | |
| nova-pro2025.12 | 0.884 | 0.865 | 0.871 | 0.868 | 0.921 | |
| claude-sonnet-42025.12 | 0.883 | 0.891 | 0.839 | 0.865 | 0.941 | |
| nova-premier2025.12 | 0.883 | 0.821 | 0.905 | 0.861 | 0.944 | |
| claude-sonnet-3.72025.12 | 0.88 | 0.861 | 0.872 | 0.866 | 0.93 | |
| mistral-large2025.12 | 0.878 | 0.804 | 0.961 | 0.876 | 0.939 | |
| llama3.3-70b2025.12 | 0.872 | 0.803 | 0.945 | 0.868 | 0.935 | |
| llama3.2-11b2025.12 | 0.867 | 0.782 | 0.972 | 0.867 | 0.899 | |
| llama4-maverick-17b2025.12 | 0.863 | 0.788 | 0.946 | 0.86 | 0.938 | |
| nova-lite2025.12 | 0.807 | 0.927 | 0.608 | 0.734 | 0.8 |