Visual Reasoning on VizWiz (Discriminability)
0.24DiscriminabilityVisual Fidelity
Evaluation Results
| Method | Links | |
|---|---|---|
| Visual FidelityModel=LLaVA2025.09 | 0.24 | |
| Prod(VF, Contr.)Model=LLaVA2025.09 | 0.168 | |
| Min(VF, Contr.)Model=LLaVA2025.09 | 0.165 | |
| SimulatabilityModel=Qwen2.52025.09 | 0.158 | |
| Avg(VF, Contr.)Model=LLaVA2025.09 | 0.141 | |
| SimulatabilityModel=GPT-4o2025.09 | 0.13 | |
| InformativenessModel=GPT-4o2025.09 | 0.113 | |
| ContrastivenessModel=GPT-4o2025.09 | 0.097 | |
| InformativenessModel=LLaVA2025.09 | 0.093 | |
| SimulatabilityModel=LLaVA2025.09 | 0.088 | |
| Min(VF, Contr.)Model=GPT-4o2025.09 | 0.084 | |
| Prod(VF, Contr.)Model=Qwen2.52025.09 | 0.082 | |
| Prod(VF, Contr.)Model=GPT-4o2025.09 | 0.082 | |
| Min(VF, Contr.)Model=Qwen2.52025.09 | 0.077 | |
| ContrastivenessModel=Qwen2.52025.09 | 0.054 | |
| Avg(VF, Contr.)Model=Qwen2.52025.09 | 0.048 | |
| Visual FidelityModel=Qwen2.52025.09 | 0.042 | |
| ContrastivenessModel=LLaVA2025.09 | 0.041 | |
| Avg(VF, Contr.)Model=GPT-4o2025.09 | 0.037 | |
| Visual FidelityModel=GPT-4o2025.09 | 0.024 | |
| PlausibilityModel=Qwen2.52025.09 | 0.02 | |
| InformativenessModel=Qwen2.52025.09 | 0.007 | |
| PlausibilityModel=LLaVA2025.09 | 0.004 | |
| PlausibilityModel=GPT-4o2025.09 | 0.002 |