AI-generated content detection on MoM (test)
93.1AccuracyCLAM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CLAMModality=Multimodal, Alignment Loss=Triplet Loss2025.11 | 93.1 | 92.5 | |
| CLAMModality=Multimodal, Alignment Loss=Cosine Loss2025.11 | 90.8 | 90.3 | |
| Poin-HierNetCategory=State of the Art Models2025.11 | 90.3 | 89.6 | |
| CLAMModality=Multimodal, Alignment Loss=Huber Loss (delta=1.0)2025.11 | 90.3 | 90.1 | |
| CLAMModality=Multimodal, Alignment Loss=MSE Loss2025.11 | 90.1 | 89.9 | |
| CLAMModality=Multimodal, Alignment Loss=None2025.11 | 88.8 | 87.5 | |
| MiOCategory=State of the Art Models2025.11 | 88.4 | 87.2 | |
| SpecTTTra-αCategory=State of the Art Models, Reference=SONICS2025.11 | 87.2 | 86.9 | |
| MERT OnlyModality=Unimodal Only2025.11 | 86.1 | 85.3 | |
| MERTModality=Unimodal Only, Configuration=No WCA2025.11 | 83.5 | 81.3 | |
| Wav2Vec2 OnlyModality=Unimodal Only2025.11 | 83.1 | 84.6 | |
| Wav2Vec2Modality=Unimodal Only, Configuration=No WCA2025.11 | 82.7 | 80.2 | |
| HUBERTModality=Unimodal Only, Configuration=No WCA2025.11 | 82 | 78.9 | |
| ASTModality=Unimodal Only, Configuration=No WCA2025.11 | 81.6 | 80.1 |