| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| HatefulMemes | Qwen2-VL-7B | AUC91.1 | 43 | 4d ago | |
| MAMI | Qwen2-VL-7B | Accuracy0.799 | 30 | 4d ago | |
| HarMeme | LLaVA-1.5-7B | Accuracy88.2 | 30 | 4d ago | |
| PrideMM | LLaVA-1.5-7B | Accuracy78.1 | 28 | 4d ago | |
| Harm-P | Qwen2-VL-7B | Accuracy91.6 | 28 | 4d ago | |
| MultiOFF | Qwen2-VL-7B | Accuracy71.1 | 27 | 4d ago | |
| HarMeme Dataset (test) | KID-VLM | Accuracy85.03 | 11 | 4d ago | |
| Tamil Memes (test) | transformer-transformer architecture | Support667 | 5 | 3d ago |
