Reward Modeling on VideoFeedback (test)
82.62Visual ScoreDDP
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DDPMethod Category=Data Pruning2026.05 | 82.62 | 75.09 | 59.26 | 51.58 | 79.28 | 69.57 | |
| DDRMethod Category=Adaptive Balancing2026.05 | 82.24 | 74.24 | 57.59 | 52.49 | 80.36 | 69.38 | |
| RLWMethod Category=Adaptive Balancing2026.05 | 82.05 | 73.63 | 56.8 | 48.8 | 78.58 | 67.97 | |
| Uncertainty WeightingMethod Category=Adaptive Balancing2026.05 | 81.47 | 74.15 | 56.7 | 49.6 | 79.15 | 68.21 | |
| High-LossMethod Category=Data Pruning2026.05 | 81.36 | 74.15 | 57.93 | 50.76 | 79.18 | 68.67 | |
| Equal WeightingMethod Category=Baseline2026.05 | 81.16 | 73.55 | 57.08 | 50.61 | 78.91 | 68.26 | |
| TracInMethod Category=Data Pruning2026.05 | 80.46 | 72.93 | 57.24 | 50.18 | 78.2 | 67.8 | |
| UncertaintyMethod Category=Data Pruning2026.05 | 80.07 | 73.81 | 55.34 | 50.81 | 77.41 | 67.49 |