Policy Steering on Hardware - PnP Cup Straightforward
0.9TPUPS
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| UPSClarification=true, Residual=true2026.02 | 0.9 | 0 | 0.1 | 0 | |
| UPSClarification=true2026.02 | 0.85 | 0.05 | 0.05 | 0.05 | |
| Forewarn2026.02 | 0.8 | 0.1 | 0.1 | 0 | |
| HG-DaggerResidual=true2026.02 | 0.5 | 0.5 | — | — | |
| EnsembleDaggerResidual=true2026.02 | 0.4 | 0.6 | — | — | |
| Base Policy2026.02 | 0.35 | 0.65 | — | — |