Video Hallucination Evaluation on VideoHallu 1.0 (test)
42.7ECTCD
Evaluation results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TCDBackbone=Video-LLaVA2026.01 | 42.7 | 48.8 | 52.9 | 40.5 | 23.8 | 32.4 | 36 | 34.4 | 36.2 | 38.1 | 29.4 | 42.2 | 50 | 40.3 | |
| Video-LLaVADecoding=Vanilla/Greedy2026.01 | 41.1 | 49.3 | 51.7 | 42.9 | 23.8 | 35.2 | 37.1 | 34.4 | 33.3 | 28.6 | 29.4 | 41.1 | 50 | 40.3 | |
| Dino-HealBackbone=Video-LLaVA2026.01 | 40 | 42.9 | 42 | 33.3 | 21.4 | 25.3 | 27.7 | 31.2 | 24.6 | 33.3 | 11.8 | 33.3 | 50 | 33 | |
| LLaVA-NeXTDecoding=Vanilla/Greedy2026.01 | 40 | 49.3 | 50 | 33.3 | 33.3 | 33.1 | 30.3 | 39.3 | 34.8 | 33.3 | 29.4 | 34.4 | 70 | 37.8 | |
| SSCDBackbone=LLaVA-NeXT2026.01 | 37.3 | 46.3 | 56.3 | 35.7 | 23.8 | 31.7 | 31.6 | 42.6 | 43.5 | 38.1 | 35.3 | 30 | 80 | 38.3 | |
| SSCDBackbone=Video-LLaVA2026.01 | 36.8 | 46.8 | 59.2 | 42.9 | 23.8 | 35.9 | 38.8 | 39.3 | 44.9 | 38.1 | 47.1 | 45.6 | 70 | 42.4 | |
| Dino-HealBackbone=LLaVA-NeXT2026.01 | 36.8 | 45.3 | 54.6 | 31 | 28.6 | 33.8 | 29.2 | 41 | 34.8 | 33.3 | 29.4 | 33.3 | 60 | 36.9 | |
| TCDBackbone=LLaVA-NeXT2026.01 | 35.7 | 47.3 | 47.7 | 35.7 | 33.3 | 31.7 | 30.7 | 41 | 31.9 | 42.9 | 47.1 | 31.1 | 80 | 36.9 | |
| MotionCDBackbone=Video-LLaVA2026.01 | 35.1 | 38.9 | 42 | 28.6 | 16.7 | 31.7 | 32.6 | 41 | 29 | 42.9 | 41.2 | 34.4 | 10 | 34.6 | |
| MotionCDBackbone=LLaVA-NeXT2026.01 | 32.4 | 40 | 50 | 33.3 | 31 | 31.7 | 29.2 | 42.6 | 29 | 33.3 | 35.3 | 27.8 | 50 | 34.5 |