Continuous Control on PEGINSERT
67.2Success RateLC-BC
Evaluation results
| Method | Links | |
|---|---|---|
| LC-BCFamily=Language-critique (BC), Supervision=Language labels2026.07 | 67.2 | |
| TD3+BCFamily=Policy-regularized, Supervision=Scalar rewards2026.07 | 58.4 | |
| LC-DPFamily=Language-critique (DP), Supervision=Language labels2026.07 | 56.8 | |
| EDPFamily=Diffusion-based, Supervision=Scalar rewards2026.07 | 56 | |
| DTFamily=Sequence-modeling, Supervision=Scalar rewards2026.07 | 40.4 | |
| CQLFamily=Value-based, Supervision=Scalar rewards2026.07 | 1.6 |