Object Hallucination Evaluation on A-OKVQA POPE (Random)
92.1AccuracySIRA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SIRABackbone=Qwen2.5-VL2026.05 | 92.1 | — | — | 91.93 | |
| MaskCDBackbone=Qwen2.5-VL2026.05 | 90.55 | — | — | 90.6 | |
| CAIBackbone=Qwen-VL2026.06 | 90.07 | — | — | 89.91 | |
| CAIBackbone=InstructBLIP2026.06 | 89.87 | — | — | 89.73 | |
| VCDBackbone=Qwen-VL2026.06 | 89.77 | — | — | 89.5 | |
| HDDModel=LLaVA-1.52025.12 | 89.5 | — | — | 89.4 | |
| HDDModel=LLaVA-NeXT2025.12 | 89.5 | — | — | 89.6 | |
| ONLYBackbone=Qwen-VL2026.06 | 89.23 | — | — | 88.78 | |
| VCDBackbone=Qwen2.5-VL2026.05 | 89.22 | — | — | 89.01 | |
| SIRABackbone=LLaVA-v1.52026.05 | 89.13 | — | — | 89.52 | |
| OPERAModel=InstructBLIP2025.12 | 89 | — | — | 89.4 | |
| HALCModel=LLaVA-NeXT2025.12 | 89 | — | — | 88.2 | |
| ICTBackbone=LLaVA-v1.52026.05 | 89 | — | — | 88.71 | |
| CAIBackbone=LLaVA-1.52026.06 | 89 | — | — | 89.03 | |
| ICTBackbone=Qwen2.5-VL2026.05 | 88.96 | — | — | 87.9 | |
| CoFi-DecModel=InstructBLIP2025.12 | 88.94 | 85.32 | — | 89.21 | |
| CoFi-DecModel=LLaVA-1.52025.12 | 88.67 | 83.63 | — | 89.38 | |
| CMPCBase Model=LLaVA 1.52025.01 | 88.53 | 90.19 | 86.47 | 88.29 | |
| SIDBackbone=Qwen2.5-VL2026.05 | 88.49 | — | — | 87.93 | |
| HALCModel=InstructBLIP2025.12 | 88.4 | — | — | 87.7 | |
| HDDModel=InstructBLIP2025.12 | 88.4 | — | — | 88.5 | |
| CoFi-DecModel=Qwen-VL2025.12 | 88.33 | 91.46 | — | 88.31 | |
| BeamModel=InstructBLIP2025.12 | 88.3 | — | — | 88.4 | |
| HALCModel=LLaVA-1.52025.12 | 88.2 | — | — | 87.4 | |
| OPERABackbone=Qwen2.5-VL2026.05 | 88.19 | — | — | 88.43 | |
| M3IDModel=Qwen-VL2025.12 | 88.13 | 92.06 | — | 87.55 | |
| OPERABackbone=LLaVA-v1.52026.05 | 88.03 | — | — | 87 | |
| DeGFModel=Qwen-VL2025.12 | 87.9 | 89.16 | — | 87.58 | |
| VCDModel=Qwen-VL2025.12 | 87.87 | 90.06 | — | 87.53 | |
| RITUALModel=Qwen-VL2025.12 | 87.73 | 92.49 | — | 87.01 | |
| VCDModel=InstructBLIP2025.12 | 87.6 | — | — | 87.7 | |
| MaskCDBackbone=LLaVA-v1.52026.05 | 87.45 | — | — | 87.55 | |
| DeGFModel=InstructBLIP2025.12 | 87.4 | 84.67 | — | 88.02 | |
| VCDModel=LLaVA-NeXT2025.12 | 87.4 | — | — | 88 | |
| DoLaBackbone=Qwen2.5-VL2026.05 | 87.4 | — | — | 86.46 | |
| VCDModel=LLaVA-1.52025.12 | 87.2 | — | — | 88.2 | |
| RITUALModel=InstructBLIP2025.12 | 87.13 | 83.92 | — | 87.71 | |
| CMPCBase Model=InstructBLIP2025.01 | 87.03 | 88.71 | 84.87 | 86.75 | |
| PAIBackbone=Qwen-VL2026.06 | 87.03 | — | — | 85.8 | |
| BeamModel=LLaVA-NeXT2025.12 | 86.6 | — | — | 85.3 | |
| RegularBackbone=Qwen-VL2026.06 | 86.53 | — | — | 85.19 | |
| SIDBackbone=LLaVA-v1.52026.05 | 86.47 | — | — | 86.52 | |
| BaseBackbone=Qwen2.5-VL2026.05 | 86.4 | — | — | 85.07 | |
| BeamModel=LLaVA-1.52025.12 | 86.3 | — | — | 87.4 | |
| OPERAModel=LLaVA-NeXT2025.12 | 86.3 | — | — | 87.2 | |
| RegularModel=Qwen-VL2025.12 | 86.27 | 90.66 | — | 85.48 | |
| ICDBase Model=LLaVA 1.52025.01 | 86.2 | 91.07 | 80.27 | 85.33 | |
| DeGFModel=LLaVA-1.52025.12 | 86.17 | 80.84 | — | 87.27 | |
| VCDBackbone=LLaVA-v1.52026.05 | 86.15 | — | — | 86.34 | |
| VCDBase Model=LLaVA 1.52025.01 | 86.15 | 85.18 | 87.53 | 86.34 | |
| ONLYBackbone=LLaVA-1.52026.06 | 86.03 | — | — | 86.93 | |
| ICDBase Model=InstructBLIP2025.01 | 85.82 | 83.8 | 88.94 | 86.29 | |
| OPERAModel=LLaVA-1.52025.12 | 85.7 | — | — | 86.9 | |
| ONLYBackbone=InstructBLIP2026.06 | 85.67 | — | — | 86.7 | |
| M3IDModel=InstructBLIP2025.12 | 85.43 | 81.77 | — | 86.23 | |
| VCDBackbone=InstructBLIP2026.06 | 85.27 | — | — | 85.72 | |
| RITUALModel=LLaVA-1.52025.12 | 85.17 | 79.79 | — | 86.4 | |
| M3IDModel=LLaVA-1.52025.12 | 84.67 | 79.25 | — | 85.97 | |
| VCDModel=InstructBLIP2025.12 | 84.2 | 80.9 | — | 85 | |
| VCDModel=LLaVA-1.52025.12 | 83.83 | 78.05 | — | 85.34 | |
| VCDBase Model=InstructBLIP2025.01 | 83.69 | 81.84 | 86.61 | 84.56 | |
| BaseBackbone=LLaVA-v1.52026.05 | 83.45 | — | — | 82.56 | |
| BaselineBase Model=LLaVA 1.52025.01 | 83.45 | 87.24 | 78.36 | 82.56 | |
| PAIBackbone=LLaVA-1.52026.06 | 83.33 | — | — | 84.85 | |
| DoLaBackbone=LLaVA-v1.52026.05 | 83.23 | — | — | 84.83 | |
| PAIBackbone=InstructBLIP2026.06 | 83.1 | — | — | 84.46 | |
| VCDBackbone=LLaVA-1.52026.06 | 82.17 | — | — | 84.14 | |
| RegularBackbone=LLaVA-1.52026.06 | 82.07 | — | — | 83.31 | |
| RegularModel=LLaVA-1.52025.12 | 81.9 | 76.63 | — | 83.53 | |
| RegularBackbone=InstructBLIP2026.06 | 81.53 | — | — | 82.86 | |
| BaselineBase Model=InstructBLIP2025.01 | 80.91 | 77.97 | 86.16 | 81.86 | |
| RegularModel=InstructBLIP2025.12 | 80.63 | 76.82 | — | 81.92 | |
| VCDModel=Qwen-VL, Decoding=VCD2026.01 | 61.83 | 63.13 | 55.6 | 59.13 | |
| SDCDModel=Qwen-VL, Decoding=SDCD2026.01 | 61.8 | 63.27 | 56.27 | 59.56 | |
| Regular DecodingModel=Qwen-VL, Decoding=Regular2026.01 | 60.87 | 62.79 | 53.33 | 57.68 |