Continuous Control on SWEEP
98.4Success RateEDP
Evaluation Results
| Method | Links | |
|---|---|---|
| EDPFamily=Diffusion-based, Supervision=Scalar rewards2026.07 | 98.4 | |
| LC-DPFamily=Language-critique (DP), Supervision=Language labels2026.07 | 96.8 | |
| LC-BCFamily=Language-critique (BC), Supervision=Language labels2026.07 | 90 | |
| DTFamily=Sequence-modeling, Supervision=Scalar rewards2026.07 | 59.2 | |
| TD3+BCFamily=Policy-regularized, Supervision=Scalar rewards2026.07 | 38.8 | |
| CQLFamily=Value-based, Supervision=Scalar rewards2026.07 | 4 |