Reward Alignment on RBM-EVAL OOD
0.96Pearson r (VOC)ACE-Brain-0.5
Evaluation results
| Method | Links | |
|---|---|---|
| ACE-Brain-0.52026.07 | 0.96 | |
| ROBOMETERTraining Data=RBM-1M data2026.03 | 0.95 | |
| Robometer-4B2026.07 | 0.94 | |
| ROBOMETERTraining Data=RoboReward Training Data2026.03 | 0.93 | |
| RoboReward-4BTraining Data=RoboReward Training Data2026.03 | 0.88 | |
| RoboReward-8BTraining Data=RoboReward Training Data2026.03 | 0.88 | |
| RoboReward-8B2026.07 | 0.88 | |
| Qwen3-VL-8B2026.07 | 0.87 | |
| RoboDopamine-8B2026.07 | 0.8 | |
| ReWiNDTraining Data=RBM-1M data2026.03 | 0.51 | |
| GVLTraining Data=Baselines2026.03 | 0.21 | |
| VLACTraining Data=Baselines2026.03 | 0.17 | |
| VLAC-8B2026.07 | 0.17 | |
| RoboDopamineTraining Data=Baselines2026.03 | 0.08 |