Input Moderation on Harmful safety datasets Average
88.33Average F1 Score (Input Moderation)MLPM
Evaluation results
| Method | Links | |
|---|---|---|
| MLPMModel architecture=N2-9B2025.02 | 88.33 | |
| Ayub & MajumdarModel architecture=N3-4B2025.02 | 87.54 | |
| MLPMModel architecture=N3-4B2025.02 | 87.52 | |
| MLPMModel architecture=N3-30B-A3B2025.02 | 87.5 | |
| Abdelnabi et al.Model architecture=N3-4B2025.02 | 86.82 | |
| Abdelnabi et al.Model architecture=N2-9B2025.02 | 86.7 | |
| Ayub & MajumdarModel architecture=N3-30B-A3B2025.02 | 86.42 | |
| Ayub & MajumdarModel architecture=N2-9B2025.02 | 86.4 | |
| Abdelnabi et al.Model architecture=N3-30B-A3B2025.02 | 86.07 |