Discriminative Object Hallucination on POPE MSCOCO (Random)
90.77AccuracyYARD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| YARDBase Model=InstructBLIP2026.05 | 90.77 | 90.05 | |
| CATCHVLM=LLaVA-1.52026.04 | 90.43 | 90.13 | |
| CATCHVLM=InstructBLIP2026.04 | 90.17 | 89.91 | |
| Object-aligned VCDVLM=LLaVA-1.52026.04 | 89.5 | 88.5 | |
| IFCDVLM=LLaVA-1.52026.04 | 89.17 | 88.47 | |
| VaLiDVLM=LLaVA-1.52026.04 | 89.03 | 88.36 | |
| YARDBase Model=LLaVA-1.5-7B2026.05 | 89.03 | 88.41 | |
| IECD2VLM=InstructBLIP2026.04 | 89 | 88.11 | |
| IECD2VLM=LLaVA-1.52026.04 | 88.07 | 86.96 | |
| VCDVLM=LLaVA-1.52026.04 | 87.73 | 87.16 | |
| OctopusVLM=LLaVA-1.52026.04 | 87.51 | 85.4 | |
| MFCDVLM=LLaVA-1.52026.04 | 87.07 | 87.73 | |
| VCDBase Model=LLaVA-1.5-7B2026.05 | 86.77 | 85.5 | |
| AVISCBase Model=LLaVA-1.5-7B2026.05 | 86.77 | 85.29 | |
| OctopusVLM=InstructBLIP2026.04 | 86.63 | 85.3 | |
| VCDBase Model=InstructBLIP2026.05 | 86.01 | 84.96 | |
| VaLiDVLM=InstructBLIP2026.04 | 85.9 | 82.98 | |
| EVASBase Model=LLaVA-1.5-7B2026.05 | 85.74 | 84.49 | |
| LLaVA-1.5-7BBase Model=LLaVA-1.5-7B2026.05 | 85.67 | 83.71 | |
| AVISCBase Model=InstructBLIP2026.05 | 85.67 | 84.31 | |
| IFCDVLM=InstructBLIP2026.04 | 85.56 | 83.75 | |
| ICDBase Model=LLaVA-1.5-7B2026.05 | 85.2 | 84.02 | |
| VCDVLM=InstructBLIP2026.04 | 84.53 | 83.68 | |
| M3IDBase Model=LLaVA-1.5-7B2026.05 | 84.5 | 82.55 | |
| ICDVLM=LLaVA-1.52026.04 | 84.23 | 82.06 | |
| ICDVLM=InstructBLIP2026.04 | 83.5 | 82.52 | |
| RegularVLM=LLaVA-1.52026.04 | 83.29 | 81.33 | |
| InstructBLIPBase Model=InstructBLIP2026.05 | 82.61 | 82.47 | |
| RegularVLM=InstructBLIP2026.04 | 80.71 | 80.41 |