CarGoal2 on Safety Gymnasium
0.65RewardCPQ
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CPQ2026.04 | 0.65 | 3.75 | |
| CPQcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.65 | 3.75 | |
| DT + SASBackbone=DT, Alignment strategy (SAS)=Lyapunov-Guided2026.04 | 0.48 | 1.15 | |
| CDTBackbone=CDT2026.04 | 0.48 | 1.25 | |
| DT + SAScost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.48 | 1.15 | |
| CDTcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.48 | 1.25 | |
| CDT + SASBackbone=CDT, Alignment strategy (SAS)=Lyapunov-Guided2026.04 | 0.42 | 0.98 | |
| CDT + SAScost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.42 | 0.98 | |
| BCQ-Lag2026.04 | 0.3 | 1.44 | |
| BCQ-Lagcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.3 | 1.44 | |
| BEAR-Lag2026.04 | 0.28 | 1.01 | |
| BEAR-Lagcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.28 | 1.01 | |
| COptiDICE2026.04 | 0.25 | 0.91 | |
| COptiDICEcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.25 | 0.91 | |
| BC-All2026.04 | 0.23 | 1.05 | |
| BC-Allcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.23 | 1.05 | |
| BC-Safe2026.04 | 0.14 | 0.51 | |
| BC-Safecost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 0.14 | 0.51 | |
| DCRL2026.04 | 0.11 | 2.51 |