Single-agent Reinforcement Learning (SARL) on Seaquest Poisoned
1,656.1Average Episode ReturnBehaviorGuard
Evaluation Results
| Method | Links | |
|---|---|---|
| BehaviorGuard2026.05 | 1,656.1 | |
| BIRD2026.05 | 1,630.8 | |
| PD2026.05 | 825.1 | |
| Original2026.05 | 141.2 |
| Method | Links | |
|---|---|---|
| BehaviorGuard2026.05 | 1,656.1 | |
| BIRD2026.05 | 1,630.8 | |
| PD2026.05 | 825.1 | |
| Original2026.05 | 141.2 |