Continuous Control on BOX-CLOSE
0.804Success RateLC-BC
Evaluation Results
| Method | Links | |
|---|---|---|
| LC-BCFamily=Language-critique (BC), Supervision=Language labels2026.07 | 0.804 | |
| EDPFamily=Diffusion-based, Supervision=Scalar rewards2026.07 | 0.716 | |
| LC-DPFamily=Language-critique (DP), Supervision=Language labels2026.07 | 0.692 | |
| TD3+BCFamily=Policy-regularized, Supervision=Scalar rewards2026.07 | 0.568 | |
| DTFamily=Sequence-modeling, Supervision=Scalar rewards2026.07 | 0.42 | |
| CQLFamily=Value-based, Supervision=Scalar rewards2026.07 | 0.1 |