Offline Reinforcement Learning on OGBench puzzle-3x3
41Average Task Success RateValue Flows
Evaluation Results
| Method | Links | |
|---|---|---|
| Value FlowsPolicy Type=Flow Policies2026.05 | 41 | |
| floqPolicy Type=Flow Policies2026.05 | 37 | |
| FlowIQN-FQLPolicy Type=Flow Policies2026.05 | 32 | |
| IQNPolicy Type=Gaussian Policies2026.05 | 31 | |
| FlowIQN-RPolicy Type=Flow Policies2026.05 | 30 | |
| FQLPolicy Type=Gaussian Policies2026.05 | 29 | |
| ReBRACPolicy Type=Gaussian Policies2026.05 | 21 | |
| CODACPolicy Type=Gaussian Policies2026.05 | 20 | |
| BCPolicy Type=Gaussian Policies2026.05 | 2 |