Hallucination Examination across VidHalluc, VideoHallucer, and EventHallusion
81.2VidHalluc ScoreGPT-4o
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-4oBackbone=Proprietary2025.12 | 81.2 | 53.3 | 91.9 | 75.5 | |
| SEASONBackbone=QWEN2.5-VL-7B, Training-free=true2025.12 | 78.7 | 54.8 | 66 | 66.5 | |
| SEASONBackbone=LLaVA-OV-7B, Training-free=true2025.12 | 75.4 | 48.5 | 69.2 | 64.3 | |
| TCDBackbone=QWEN2.5-VL-7B, Training-free=true2025.12 | 74.8 | 52.8 | 64.8 | 64.1 | |
| DINO-HEALBackbone=QWEN2.5-VL-7B, Training-free=true2025.12 | 73.7 | 53.3 | 63.6 | 63.5 | |
| QWEN2.5-VL-7BBackbone=QWEN2.5-VL-7B2025.12 | 73.4 | 53.1 | 63.3 | 63.3 | |
| ArrowRLBackbone=QWEN2.5-VL-7B, Training-free=false2025.12 | 73.2 | 54.5 | 69 | 65.5 | |
| Gemini 1.5 ProBackbone=Proprietary2025.12 | 72.8 | 37.8 | 80.4 | 63.8 | |
| TCDBackbone=LLaVA-OV-7B, Training-free=true2025.12 | 71.4 | 48.1 | 68.5 | 62.6 | |
| LLaVA-OV-7BBackbone=LLaVA-OV-7B2025.12 | 70.3 | 46.4 | 60.2 | 59 | |
| DINO-HEALBackbone=LLaVA-OV-7B, Training-free=true2025.12 | 70.2 | 46.9 | 60.2 | 59.1 | |
| SEASONBackbone=LLaVA-Video-7B, Training-free=true2025.12 | 66.5 | 51.4 | 67 | 61.6 | |
| TCDBackbone=LLaVA-Video-7B, Training-free=true2025.12 | 64.9 | 50.1 | 64.3 | 59.8 | |
| TPOBackbone=LLaVA-Video-7B, Training-free=false2025.12 | 64.4 | 51.3 | 63.3 | 59.7 | |
| RRPOBackbone=LLaVA-Video-7B, Training-free=false2025.12 | 63.9 | 51.3 | 68 | 61.1 | |
| LLaVA-Video-7BBackbone=LLaVA-Video-7B2025.12 | 63.8 | 51.4 | 63.6 | 59.6 | |
| DINO-HEALBackbone=LLaVA-Video-7B, Training-free=true2025.12 | 63.8 | 51.3 | 64.3 | 59.8 |