Goal-conditioned Reinforcement Learning (G1_speed) on hopper medium-expert v2
2.455Goal-conditioned ReturnFrozen
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Frozen2026.04 | 2.455 | 0 | |
| Additive2026.04 | 2.311 | 392.886 | |
| KL-Reg2026.04 | 2.282 | 27.656 | |
| PoE2026.04 | 2.273 | 29.407 | |
| Prior Only2026.04 | 2.26 | 1,560.719 |