MIRAGE
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| AI-Generated Image Detection | Mirage (test) | Human Overall Accuracy99.18 | 14 | |
| Coarse-level Multimodal Misinformation Detection | MiRAGe News | Accuracy80.2 | 14 | |
| Medical Question Answering | MIRAGE (test) | MMLU-Med89.44 | 12 | |
| Biomedical Retrieval-Augmented Generation | Mirage | MMLU-med Accuracy87.24 | 10 | |
| Flicker-banding and Moire Removal | MIRAGE cropped (test) | SSIM0.7354 | 9 | |
| GUI Agent Attack Success Rate Evaluation | MIRAGE (1,111-sample main set) | FB Success Rate41 | 5 | |
| Multi-modal Forgery Detection | MiRAGe | Accuracy53.92 | 5 | |
| Binary forgery detection | MiRAGe | Accuracy56.99 | 5 | |
| Multi-choice | MIRAGE | Accuracy58.3 | 2 | |
| Dataset Diversity and Coverage Evaluation | MIRAGE 3-app overlap | Goal-Text Entropy0.918 | 1 | |
| Dataset Diversity and Coverage Evaluation | MIRAGE matched-n | Goal-text Entropy0.927 | 1 | |
| Dataset Diversity and Coverage Evaluation | MIRAGE full | Goal-text Entropy0.933 | 1 | |
| Data source relevance classification | MIRAGE (test) | Accuracy86.63 | 1 |
Showing 13 of 13 rows