| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Button2 | Safety Gymnasium | Reward9.59 | 16 | |
| Button1 | Safety Gymnasium | Reward7.65 | 16 | |
| Goal2 | Safety Gymnasium | Reward11.48 | 16 | |
| Goal1 | Safety Gymnasium | Reward7.9 | 16 | |
| Swimmer Velocity | Safety Gymnasium level-2 | Safe Reward94 | 12 | |
| Hopper Velocity | Safety Gymnasium level-2 | Safe Reward1,300 | 12 | |
| Car Goal | Safety Gymnasium level-2 | Safe Reward1.6 | 12 | |
| Car Circle | Safety Gymnasium level-2 | Safe Reward11 | 12 | |
| Car Push | Safety Gymnasium level-2 | Safe Reward0.38 | 12 | |
| Point Goal | Safety Gymnasium level-2 | Safe Reward2.2 | 12 | |
| Point Button | Safety Gymnasium level-2 | Safe Reward1.3 | 12 | |
| Point Push | Safety Gymnasium level-2 | Safe Reward0.33 | 12 | |
| Button navigation | Safety Gymnasium Button2 v0 (test) | Success Rate100 | 8 | |
| Button navigation | Safety Gymnasium Button1 v0 (test) | Success Rate100 | 8 | |
| Goal achievement | Safety Gymnasium Goal2 v0 (test) | Success Rate1 | 8 | |
| Goal achievement | Safety Gymnasium Goal1 v0 (test) | Success Rate100 | 8 | |
| Continuous Control | Safety-Gymnasium SafetyAntVelocity (generalization) | Episodic Reward45.5 | 6 | |
| Continuous Control | Safety-Gymnasium SafetyCarGoal1 (generalization) | Episodic Reward26.8 | 6 | |
| Continuous Control | Safety-Gymnasium SafetyPointGoal1 (generalization) | Episodic Reward24.5 | 6 | |
| Safe Reinforcement Learning | Safety Gymnasium Button, threshold=40 | Return9.01 | 4 | |
| Safe Reinforcement Learning | Safety Gymnasium Push, threshold=35 | Return4.77 | 4 | |
| Safe Reinforcement Learning | Safety Gymnasium Goal, threshold=20 | Return24.26 | 4 | |
| Safe Reinforcement Learning | Safety Gymnasium Humanoid threshold=5 | Return5,948 | 4 | |
| Safe Reinforcement Learning | Safety Gymnasium Walker2d, threshold=5 | Return3,057 | 4 | |
| Safe Reinforcement Learning | Safety Gymnasium HalfCheetah, threshold=5 | Return2,619 | 4 |