Object Hallucination Evaluation on AOKVQA POPE (Random, Popular, and Adversarial)
82.23AccuracySVC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SVCBase MLLM=LLaVA-1.52026.03 | 82.23 | 83.82 | |
| SVCBase MLLM=Shikra2026.03 | 81.87 | 82.18 | |
| ONLYBase MLLM=Shikra2026.03 | 81.64 | 81.78 | |
| VISTABase MLLM=LLaVA-1.52026.03 | 81.23 | 81.65 | |
| ONLYBase MLLM=InstructBLIP2026.03 | 80.76 | 82.58 | |
| ONLYBase MLLM=LLaVA-1.52026.03 | 80.55 | 82.24 | |
| SVCBase MLLM=InstructBLIP2026.03 | 80.24 | 82.48 | |
| VISTABase MLLM=Shikra2026.03 | 80.12 | 81.24 | |
| VISTABase MLLM=InstructBLIP2026.03 | 80.09 | 82.3 | |
| PAIBase MLLM=Shikra2026.03 | 79.97 | 81.07 | |
| PAIBase MLLM=InstructBLIP2026.03 | 79.04 | 81.77 | |
| VCDBase MLLM=InstructBLIP2026.03 | 78.95 | 81.52 | |
| VanillaBase MLLM=InstructBLIP2026.03 | 78.92 | 81.51 | |
| VanillaBase MLLM=Shikra2026.03 | 78.67 | 80.03 | |
| PAIBase MLLM=LLaVA-1.52026.03 | 78.65 | 79.82 | |
| VCDBase MLLM=Shikra2026.03 | 78.52 | 80.24 | |
| VanillaBase MLLM=LLaVA-1.52026.03 | 77.23 | 80.62 | |
| VCDBase MLLM=LLaVA-1.52026.03 | 76.29 | 79.99 | |
| PAIBase MLLM=MiniGPT-42026.03 | 73.2 | 73.22 | |
| SVCBase MLLM=MiniGPT-42026.03 | 73.13 | 73.69 | |
| VanillaBase MLLM=MiniGPT-42026.03 | 72.8 | 72.99 | |
| ONLYBase MLLM=MiniGPT-42026.03 | 72.54 | 72.56 | |
| VISTABase MLLM=MiniGPT-42026.03 | 71.82 | 72.34 | |
| VCDBase MLLM=MiniGPT-42026.03 | 71.27 | 71.93 |