Semantic Attribution on Ego4D-M (test)
62.03Average AccuracyMisFormer
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MisFormerSubset evaluation (ChatGPT baseline fair comparison)=false2025.11 | 62.03 | 56.24 | 58.4 | 55.22 | 65.66 | 57.25 | |
| MisFormerSubset evaluation (ChatGPT baseline fair comparison)=true2025.11 | 59.37 | 55.41 | 55.44 | 53.03 | 63.33 | 57.79 | |
| ChatGPT2025.11 | 52.45 | 50.95 | 53.42 | 50.2 | 51.48 | 51.7 | |
| LLaVA-Vid2025.11 | 42.83 | 40.12 | 48.1 | 47.95 | 37.56 | 32.29 | |
| Vid-Chat2025.11 | 36.47 | 34.36 | 36.73 | 36.84 | 36.21 | 31.88 |