Video Preference Alignment on MonteBench
66.36Alignment Accuracy (w/ties)GT-SVJ
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GT-SVJHuman Dataset Size=30.4K, Backbone Size=2B2026.02 | 66.36 | 77.76 | |
| VisionRewardHuman Dataset Size=2000K, Backbone Size=19B2026.02 | 64 | 72.1 | |
| GT-SVJ (p = 0)Human Dataset Size=30.4K, Backbone Size=2B2026.02 | 62.79 | 76.12 | |
| GT-SVJ (No DM)Human Dataset Size=30.4K, Backbone Size=2B2026.02 | 61.82 | 73.59 | |
| VideoRewardHuman Dataset Size=182K, Backbone Size=2B2026.02 | 54.2 | 62.25 | |
| VideoScoreHuman Dataset Size=37.6K, Backbone Size=8B2026.02 | 49.1 | 54.9 |