Policy Optimization on Traffic
16.91True OutcomeORPO
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ORPOEnvironment=Traffic2026.04 | 16.91 | 3.41 | -19,600 | -0.69 | -0.83 | 0.0004 | |
| ORPO2026.04 | 16.91 | 3.41 | -19,600 | -0.69 | -0.83 | 0.0004 | |
| Linear Max-MinEnvironment=Traffic2026.04 | 16.46 | 2.4 | -11,900 | 0.2 | -0.12 | 0 | |
| Linear Max-Min2026.04 | 16.46 | 2.4 | -11,900 | 0.2 | -0.12 | 0 | |
| Max-MinEnvironment=Traffic2026.04 | 12.7 | 3.63 | -268.31 | -0.06 | -0.06 | 0 | |
| Max-Min2026.04 | 12.7 | 3.63 | -268.31 | -0.06 | -0.06 | 0 | |
| ORPO*Environment=Traffic2026.04 | 10.26 | 1.35 | -13,500 | -0.44 | -0.45 | 0.0002 | |
| ORPO*2026.04 | 10.26 | 1.35 | -13,500 | -0.44 | -0.45 | 0.0002 |