Continuous Control on Walker-Walk ID
4.68Paired Delta (Episode Returns)HarderPairs
Evaluation results
| Method | Links | |
|---|---|---|
| HarderPairsBlock=First2026.04 | 4.68 | |
| V2 (naive preference)Block=Second2026.04 | 4.24 | |
| HarderPairsBlock=Second2026.04 | 3.64 | |
| V2 (naive preference)Block=First2026.04 | 1.31 |