Fact Verification on InfoTabs (held-out)
79.5AccuracyGPT-4o
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4oResolution=UNK2025.09 | 79.5 | |
| OpenAI-o4-miniResolution=UNK2025.09 | 78.3 | |
| Gemini-2.5-ProResolution=UNK2025.09 | 77.15 | |
| QVQ-72B-PreviewResolution=Dyn.2025.09 | 74.64 | |
| Qwen2-VL-72BResolution=Dyn.2025.09 | 72.82 | |
| Claude-3.5-SonnetResolution=UNK2025.09 | 70.3 | |
| Table-LLaVA 13BResolution=336, Trained on dataset=true2025.09 | 66.91 | |
| Table-LLaVA 7BResolution=336, Trained on dataset=true2025.09 | 65.26 | |
| Qwen2-VL-7B-Table-R1Resolution=Dyn.2025.09 | 62.8 | |
| Qwen2-VL-7B-SFTResolution=Dyn.2025.09 | 57.6 | |
| Qwen2-VL-7B-InsResolution=Dyn2025.09 | 46.56 | |
| Qwen2-VL-2B-Table-R1Resolution=Dyn.2025.09 | 43.6 | |
| Qwen2-VL-7B-GRPOResolution=Dyn.2025.09 | 41.6 | |
| Qwen2-VL-2B-SFTResolution=Dyn.2025.09 | 37.4 | |
| Qwen2-VL-7BResolution=Dyn.2025.09 | 32.8 | |
| Qwen2-VL-2B-GRPOResolution=Dyn.2025.09 | 29 | |
| LLaVA-v1.5 7BResolution=3362025.09 | 28.31 | |
| mPLUG-Owl2 7BResolution=4482025.09 | 26.19 | |
| DeepSeek-VL2 4.5BResolution=Dyn.2025.09 | 24.79 | |
| Qwen2-VL-2BResolution=Dyn.2025.09 | 24.6 | |
| Monkey 7BResolution=8962025.09 | 22.11 |