Hallucination Evaluation on Style-CHAIR Game
19.3CHAIRi ScoreDola
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DolaModel=InstructBLIP2025.08 | 19.3 | 48.3 | |
| DolaModel=LLaVA-1.52025.08 | 14.1 | 43 | |
| DolaModel=MiniGPT-42025.08 | 12.3 | 34.3 | |
| AGLAModel=MiniGPT-42025.08 | 11.2 | 43.2 | |
| OPERAModel=MiniGPT-42025.08 | 10.9 | 34.7 | |
| OPERAModel=InstructBLIP2025.08 | 10.7 | 35 | |
| BeamModel=MiniGPT-42025.08 | 10.7 | 34.3 | |
| GreedyModel=MiniGPT-42025.08 | 9.8 | 33 | |
| BeamModel=InstructBLIP2025.08 | 9.7 | 34 | |
| DecoModel=LLaVA-1.62025.08 | 9.7 | 24.3 | |
| AGLAModel=InstructBLIP2025.08 | 9.6 | 36 | |
| OPERAModel=LLaVA-1.52025.08 | 9.6 | 33 | |
| BeamModel=LLaVA-1.62025.08 | 9.6 | 24.3 | |
| GreedyModel=InstructBLIP2025.08 | 9.4 | 36.7 | |
| BeamModel=LLaVA-1.52025.08 | 9.3 | 32.3 | |
| AGLAModel=LLaVA-1.52025.08 | 9.3 | 31.3 | |
| AGLAModel=LLaVA-1.62025.08 | 9.2 | 23.3 | |
| DecoModel=MiniGPT-42025.08 | 9.2 | 29.3 | |
| GreedyModel=LLaVA-1.52025.08 | 9.1 | 31.7 | |
| DecoModel=InstructBLIP2025.08 | 8.9 | 28 | |
| DecoModel=LLaVA-1.52025.08 | 8.9 | 30 | |
| SAVERModel=LLaVA-1.52025.08 | 8.8 | 29.7 | |
| GreedyModel=LLaVA-1.62025.08 | 8.8 | 22 | |
| SAVERModel=MiniGPT-42025.08 | 8.8 | 28.7 | |
| OPERAModel=LLaVA-1.62025.08 | 8.6 | 21.5 | |
| DolaModel=LLaVA-1.62025.08 | 8.1 | 18.3 | |
| SAVERModel=InstructBLIP2025.08 | 8 | 30 | |
| SAVERModel=LLaVA-1.62025.08 | 7.4 | 16.3 |