Reinforcement Learning on Reacher Easy OOD
158.5Score (mass×0.7)HaM-World
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| HaM-Worldprotocol=zero-shot2026.05 | 158.5 | 158.8 | 141.7 | 139.7 | 148.8 | 151.8 | 149.9 | |
| TD-MPC2protocol=zero-shot2026.05 | 135.3 | 130.1 | 131.8 | 125.1 | 131.5 | 138.1 | 132 | |
| SACprotocol=zero-shot2026.05 | 98.5 | 98 | 86 | 83 | 88.6 | 100.2 | 92.4 | |
| PPOprotocol=zero-shot2026.05 | 11.7 | 11.8 | 14.6 | 13.9 | 12.7 | 12.9 | 12.9 | |
| DreamerV3protocol=zero-shot2026.05 | 8.7 | 11.4 | 9.3 | 8.1 | 8.9 | 9.4 | 9.3 |