| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Reward rollout alignment | 10-task benchmark T1: Folding Shorts | Rollout ρ0.889 | 8 | |
| Reward prediction | 10-task benchmark S1 classic | Demo L (MSE)0.006 | 8 | |
| Reward rollout alignment | 10-task benchmark T2: Cleaning Whiteboard | Rollout ρ0.667 | 7 |