Multi-modal Stance Detection on MWTWT
66.36Macro F1 (CA)TMPT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TMPTModality=Multi-modal, Zero-shot=true2026.04 | 66.36 | 66.39 | 66.32 | 61.56 | |
| MM-StanceDetModality=Multi-modal, Zero-shot=true2026.04 | 65.24 | 67.03 | 68.57 | 57.86 | |
| TASTEModality=Multi-modal, Zero-shot=true2026.04 | 65.22 | 63.48 | 65.91 | 62.77 | |
| BERTModality=Textual, Zero-shot=true2026.04 | 63.55 | 61.3 | 59.18 | 52.89 | |
| BridgeTowerModality=Multi-modal, Zero-shot=true2026.04 | 63.51 | 61.82 | 64.93 | 60.11 | |
| CLIPModality=Multi-modal, Zero-shot=true2026.04 | 61.08 | 55.67 | 63.8 | 60.06 | |
| LKI-BARTModality=Textual, Zero-shot=true2026.04 | 60.37 | 62.85 | 64.2 | 55.78 | |
| KEBERTModality=Textual, Zero-shot=true2026.04 | 59.7 | 62.56 | 63.92 | 55.53 | |
| RoBERTaModality=Textual, Zero-shot=true2026.04 | 59.22 | 59.22 | 64.86 | 57.46 | |
| BERT+ViTModality=Multi-modal, Zero-shot=true2026.04 | 59.21 | 59.3 | 65.04 | 59.28 | |
| MV-DebateModality=Textual, Zero-shot=true2026.04 | 57.81 | 60.67 | 66.15 | 69.12 | |
| GPT-4 + CoTModality=Textual, Zero-shot=true2026.04 | 57.52 | 60.85 | 65.91 | 69.3 | |
| GPT-4Modality=Textual, Zero-shot=true2026.04 | 57.19 | 60.56 | 65.63 | 69.01 | |
| GPT-4 VisionModality=Multi-modal, Zero-shot=true2026.04 | 42.23 | 45.92 | 54.59 | 53.19 | |
| Qwen-VLModality=Multi-modal, Zero-shot=true2026.04 | 38.57 | 43.36 | 47.82 | 41.01 | |
| ViLTModality=Multi-modal, Zero-shot=true2026.04 | 38.33 | 46 | 55.01 | 48.55 | |
| LLaMA2Modality=Textual, Zero-shot=true2026.04 | 32.47 | 38.37 | 48.08 | 46.13 | |
| SwinTModality=Visual, Zero-shot=true2026.04 | 28.53 | 28.5 | 35.87 | 34.33 | |
| ViTModality=Visual, Zero-shot=true2026.04 | 24.59 | 28.18 | 34.06 | 33.4 | |
| ResNetModality=Visual, Zero-shot=true2026.04 | 23.01 | 24.11 | 25.21 | 25.27 |