Edit Inspectors Questions on UltraEdit
63AccuracyGPT-4
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GPT-42025.06 | 63 | 48.5 | 53.4 | 42.9 | 55.6 | |
| Qwen2.5 VL2025.06 | 62.1 | 58.2 | 43.8 | 50 | 54.4 | |
| GPT-4 Turbo2025.06 | 57.5 | 37.1 | 48.5 | 51 | 57.5 | |
| InternVL32025.06 | 54.4 | 61.9 | 46.2 | 41 | 44.4 | |
| LLaVATraining Protocol=Supervised2025.06 | 54.3 | — | — | 55.6 | 48.1 | |
| LLaVA2025.06 | 52.6 | — | — | 52.7 | 50 | |
| GPT-4o2025.06 | 51.8 | 41.6 | 46.4 | 56.3 | 55.6 |