Hallucination Detection on EventHallusion
75AccuracyMACD
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MACDBackbone=Qwen3-VL-2B2026.02 | 75 | 76 | 97 | 85 | |
| GPT-4o2025.12 | 73.3 | — | — | — | |
| MACDBackbone=Qwen2-VL-7B2026.02 | 73 | 86 | 75 | 80 | |
| SIDBackbone=Qwen3-VL-2B2026.02 | 72 | 79 | 72 | 78 | |
| VCDBackbone=Qwen3-VL-2B2026.02 | 72 | 75 | 91 | 82 | |
| BaselineBackbone=Qwen2-VL-7B2026.02 | 72 | 84 | 75 | 79 | |
| VCDBackbone=Qwen2-VL-7B2026.02 | 72 | 82 | 78 | 80 | |
| MACDBackbone=Qwen2.5-VL-3B2026.02 | 71 | 80 | 78 | 79 | |
| SIDBackbone=Qwen2-VL-7B2026.02 | 66 | 81 | 69 | 75 | |
| SIDBackbone=Qwen2.5-VL-3B2026.02 | 65 | 76 | 74 | 75 | |
| BaselineBackbone=Qwen2.5-VL-3B2026.02 | 62 | 76 | 69 | 72 | |
| VCDBackbone=Qwen2.5-VL-3B2026.02 | 62 | 74 | 73 | 73 | |
| DNA-Train-7B2025.12 | 61.3 | — | — | — | |
| MACDBackbone=Qwen2.5-VL-7B2026.02 | 61 | 82 | 57 | 67 | |
| BaselineBackbone=Qwen3-VL-2B2026.02 | 60 | 76 | 61 | 68 | |
| VCDBackbone=Qwen2.5-VL-7B2026.02 | 55 | 78 | 50 | 61 | |
| MACDBackbone=Qwen2-VL-2B2026.02 | 51 | 71 | 53 | 61 | |
| VCDBackbone=Qwen2-VL-2B2026.02 | 49 | 66 | 58 | 62 | |
| SIDBackbone=Qwen2-VL-2B2026.02 | 46 | 75 | 36 | 49 | |
| BaselineBackbone=Qwen2.5-VL-7B2026.02 | 44 | 75 | 31 | 44 | |
| MACDBackbone=InternVL3-8B2026.02 | 44 | 67 | 42 | 42 | |
| SIDBackbone=InternVL3-8B2026.02 | 42 | 67 | 36 | 47 | |
| SIDBackbone=Qwen2.5-VL-7B2026.02 | 38 | 76 | 18 | 29 | |
| VCDBackbone=InternVL3-8B2026.02 | 37 | 64 | 25 | 35 | |
| BaselineBackbone=Qwen2-VL-2B2026.02 | 36 | 61 | 26 | 36 | |
| BaselineBackbone=InternVL3-8B2026.02 | 35 | 60 | 25 | 35 | |
| Qwen2.5-VL-7B2025.12 | 33.5 | — | — | — | |
| Video-LLaVA-7B2025.12 | 29.7 | — | — | — | |
| VideoChat2-HD2025.12 | 20 | — | — | — | |
| LLaVA-Next-Video2025.12 | 12.1 | — | — | — |