Object Hallucination Evaluation on MS-COCO POPE Random
92.36AccuracyAVISC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| AVISCModel=Qwen2.5-VL-7B2024.05 | 92.36 | 91.5 | — | — | |
| AVISCModel=LLaVA-OneVision (Qwen2-7B)2024.05 | 91.46 | 90.84 | — | — | |
| SIRABackbone=Qwen2.5-VL2026.05 | 91.13 | 90.49 | — | — | |
| SchroMindModel=LLaVA-1.5-7B2026.02 | 90.86 | 89.11 | — | — | |
| M3IDModel=Qwen2.5-VL-7B2024.05 | 90.77 | 88.57 | — | — | |
| VCDModel=LLaVA-OneVision (Qwen2-7B)2024.05 | 90.57 | 89.71 | — | — | |
| VCDModel=Qwen2.5-VL-7B2024.05 | 90.47 | 89.37 | — | — | |
| CoFi-DecModel=LLaVA-1.52025.12 | 90.33 | 89.38 | 89.05 | — | |
| MESAModel=LLaVA-v1.52026.04 | 90.27 | 90.2 | 95.81 | 85.2 | |
| CoFi-DecModel=InstructBLIP2025.12 | 90.12 | 91.38 | 91.23 | — | |
| CoFi-DecModel=Qwen-VL2025.12 | 90.11 | 88.92 | 96.87 | — | |
| MaskCDBackbone=Qwen2.5-VL2026.05 | 90.05 | 89.6 | — | — | |
| M3IDModel=LLaVA-OneVision (Qwen2-7B)2024.05 | 89.87 | 88.91 | — | — | |
| M3IDModel=Qwen-VL2025.12 | 89.83 | 89.17 | 95.44 | — | |
| DeGFModel=Qwen-VL2025.12 | 89.73 | 89.31 | 93.19 | — | |
| RITUALModel=Qwen-VL2025.12 | 89.47 | 88.62 | 96.32 | — | |
| DeGFModel=InstructBLIP2025.12 | 89.3 | 89.1 | 90.68 | — | |
| SIRABackbone=LLaVA-v1.52026.05 | 89.27 | 89.18 | — | — | |
| DeGFModel=LLaVA-1.52025.12 | 89.23 | 89.11 | 90.17 | — | |
| CMPCBase Model=LLaVA 1.52025.01 | 89.23 | 88.68 | 93.5 | 84.33 | |
| MoDModel=LLaVA-v1.52025.05 | 89.2 | 89.1 | 90 | 88.2 | |
| OPERAModel=LLaVA-1.5-7B2026.02 | 89.2 | 88.81 | — | — | |
| OPERABackbone=LLaVA-v1.52026.05 | 89.2 | 88.81 | — | — | |
| ICTBackbone=LLaVA-v1.52026.05 | 89.18 | 88.48 | — | — | |
| VDD-NoneBackbone=LLaVA-1.5-7B, Decoding Strategy=VDD-None2026.05 | 89.14 | 89.53 | — | — | |
| BRACSBackbone=LLaVA-1.5-7B, Decoding Strategy=BRACS2026.05 | 89.11 | 89.17 | — | — | |
| ICTModel=LLaVA-1.5-7B2026.02 | 89.1 | 88.48 | — | — | |
| NulluModel=LLaVA-v1.52026.04 | 89.03 | 88.52 | 92.83 | 84.6 | |
| PAIBackbone=LLaVA-1.5-7B, Decoding Strategy=PAI2026.05 | 88.97 | 89.48 | — | — | |
| ICTModel=LLaVA-v1.52026.04 | 88.96 | 87.85 | 96.87 | 80.4 | |
| RITUALModel=LLaVA-1.52025.12 | 88.87 | 88.81 | 89.23 | — | |
| baseModel=Qwen2.5-VL-7B2024.05 | 88.84 | 87.82 | — | — | |
| RITUALModel=InstructBLIP2025.12 | 88.83 | 88.6 | 90.48 | — | |
| MESAModel=Qwen-VL2026.04 | 88.83 | 87.11 | 97.29 | 78.87 | |
| VCDModel=Qwen-VL2025.12 | 88.8 | 88.11 | 93.89 | — | |
| VCDModel=Qwen-VL2026.04 | 88.63 | 87.81 | 94.64 | 81.91 | |
| RVEBase Model=LLaVA-1.5-7B2026.05 | 88.63 | 87.51 | — | — | |
| VCDBackbone=Qwen2.5-VL2026.05 | 88.63 | 87.81 | — | — | |
| baseModel=LLaVA-OneVision (Qwen2-7B)2024.05 | 88.6 | 87.34 | — | — | |
| MaskCDBackbone=LLaVA-v1.52026.05 | 88.55 | 88.3 | — | — | |
| AVISCModel=LLaVA-1.5 (13B)2024.05 | 88.4 | 88.77 | 86.05 | 91.67 | |
| VCDBase Model=LLaVA-1.5-7B2026.05 | 88.37 | 87.91 | — | — | |
| VAFBase Model=LLaVA-1.5-7B2026.05 | 88.3 | 87.29 | — | — | |
| NulluModel=Qwen-VL2026.04 | 88.26 | 86.13 | 91.47 | 81.4 | |
| BRACSBackbone=Qwen-VL-Chat, Decoding Strategy=BRACS2026.05 | 88.25 | 87.39 | — | — | |
| VDD-NoneBackbone=Qwen-VL-Chat, Decoding Strategy=VDD-None2026.05 | 88.18 | 87.31 | — | — | |
| ICTModel=Qwen-VL2026.04 | 88.06 | 86.55 | 96.9 | 78.11 | |
| SIDBackbone=LLaVA-v1.52026.05 | 88.05 | 87.34 | — | — | |
| VTIModel=Qwen-VL2026.04 | 87.97 | 86.79 | 96.19 | 79.07 | |
| PAIBackbone=Qwen-VL-Chat, Decoding Strategy=PAI2026.05 | 87.97 | 87.12 | — | — | |
| AVISCModel=LLaVA-1.5-7B2026.02 | 87.93 | 87.88 | — | — | |
| SIDBackbone=Qwen2.5-VL2026.05 | 87.9 | 86.73 | — | — | |
| VCDModel=LLaVA-1.5-7B2026.02 | 87.73 | 87.16 | — | — | |
| VCDModel=LLaVA-v1.52026.04 | 87.73 | 87.16 | 91.42 | 83.28 | |
| VCDBackbone=LLaVA-v1.52026.05 | 87.73 | 87.16 | — | — | |
| VCDBase Model=LLaVA 1.52025.01 | 87.73 | 87.16 | 91.42 | 83.28 | |
| VTIModel=LLaVA-v1.52026.04 | 87.7 | 86.45 | 96.24 | 78.47 | |
| PNDModel=InstructBLIP-7B2026.04 | 87.63 | 86.73 | 93.52 | 80.86 | |
| ICTBackbone=Qwen2.5-VL2026.05 | 87.53 | 85.84 | — | — | |
| M3IDModel=LLaVA-1.52025.12 | 87.5 | 87.52 | 87.38 | — | |
| RegularModel=Qwen-VL2025.12 | 87.43 | 86.48 | 93.56 | — | |
| VAFModel=LLaVA-v1.52026.04 | 87.36 | 85.94 | 96.82 | 77.26 | |
| PNDModel=LLaVA1.5-7B2026.04 | 87.33 | 93.41 | 98.29 | 88.99 | |
| OPERABackbone=Qwen2.5-VL2026.05 | 87.31 | 86.92 | — | — | |
| VanillaModel=Qwen-VL2026.04 | 87.23 | 85.5 | 91.7 | 80.06 | |
| AGLAModel=LLaVA1.5-7B2026.04 | 87.12 | 92.44 | 96.42 | 88.78 | |
| VanillaBase Model=LLaVA-1.5-7B2026.05 | 87.07 | 85.5 | — | — | |
| VCDModel=LLaVA-1.52025.12 | 87 | 87.15 | 86.13 | — | |
| CMPCBase Model=InstructBLIP2025.01 | 86.96 | 85.83 | 93.97 | 79 | |
| SPINBackbone=LLaVA-1.5-7B, Decoding Strategy=SPIN2026.05 | 86.84 | 85.66 | — | — | |
| VCDModel=LLaVA1.5-7B2026.04 | 86.81 | 90.43 | 96.11 | 85.38 | |
| VAFModel=Qwen-VL2026.04 | 86.8 | 84.98 | 98.5 | 74.73 | |
| M3IDModel=InstructBLIP2025.12 | 86.67 | 86.41 | 88.09 | — | |
| MoDModel=LLaVA-NEXT2025.05 | 86.6 | 84.8 | 97.7 | 74.9 | |
| VCDBackbone=Qwen-VL-Chat, Decoding Strategy=VCD2026.05 | 86.56 | 85.28 | — | — | |
| DoLaBackbone=Qwen2.5-VL2026.05 | 86.53 | 84.44 | — | — | |
| VCDModel=Qwen-VL, Decoding=VCD2026.01 | 86.47 | 84.7 | 95.87 | 75.87 | |
| ICDBase Model=InstructBLIP2025.01 | 86.43 | 85.61 | 92.01 | 80.73 | |
| GreedyBackbone=LLaVA-1.5-7B, Decoding Strategy=Greedy2026.05 | 86.32 | 85 | — | — | |
| AGLAModel=InstructBLIP-7B2026.04 | 86.31 | 85.56 | 92.32 | 79.72 | |
| SPINBackbone=Qwen-VL-Chat, Decoding Strategy=SPIN2026.05 | 86.29 | 84.82 | — | — | |
| VCDModel=InstructBLIP2025.12 | 86.23 | 85.88 | 88.14 | — | |
| VCDBackbone=LLaVA-1.5-7B, Decoding Strategy=VCD2026.05 | 86.22 | 85.23 | — | — | |
| GreedyBackbone=Qwen-VL-Chat, Decoding Strategy=Greedy2026.05 | 86.19 | 84.7 | — | — | |
| M3IDModel=LLaVA-v1.52025.05 | 86.1 | 85 | 83.2 | 86.8 | |
| MoDModel=Qwen-VL2025.05 | 86 | 84.1 | 97 | 74.3 | |
| VCDModel=LLaVA-NEXT2025.05 | 86 | 84.3 | 96.5 | 74.8 | |
| M3IDModel=LLaVA-1.5-7B2026.02 | 86 | 86 | — | — | |
| DoLaBackbone=LLaVA-v1.52026.05 | 85.97 | 86.14 | — | — | |
| VAFModel=LLaVA1.5-7B2026.04 | 85.68 | 91.87 | 95.33 | 88.65 | |
| VanillaModel=LLaVA-v1.52026.04 | 85.53 | 88.8 | 86.77 | 90.93 | |
| VCDModel=Qwen-VL2025.05 | 85.5 | 83.6 | 96 | 71.1 | |
| M3IDModel=LLaVA-NEXT2025.05 | 85.5 | 83.6 | 96.3 | 73.9 | |
| M3IDModel=Qwen-VL2025.05 | 85.3 | 83.4 | 95.1 | 74.2 | |
| ICDModel=LLaVA-v1.52026.04 | 85.23 | 83.52 | 94.44 | 74.86 | |
| BaseBackbone=Qwen2.5-VL2026.05 | 85.23 | 83.09 | — | — | |
| AvisCModel=LLaVA-NEXT2025.05 | 85.2 | 82.8 | 98.6 | 71.3 | |
| SDCDModel=Qwen-VL, Decoding=SDCD2026.01 | 85.17 | 83.24 | 95.67 | 73.67 | |
| ICDBase Model=LLaVA 1.52025.01 | 85.17 | 83.19 | 95.99 | 73.4 | |
| VCDModel=InstructBLIP-7B2026.04 | 85.12 | 84.18 | 88.57 | 80.2 |