QED Molecular Property Optimization on ZINC-derived difficult subset
5.45Success Rate (SR)Active-GRPO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Active-GRPO2026.07 | 5.45 | 92.99 | 5.04 | |
| RePO2026.07 | 4.68 | 95.38 | 4.46 |
| Method | Links | |||
|---|---|---|---|---|
| Active-GRPO2026.07 | 5.45 | 92.99 | 5.04 | |
| RePO2026.07 | 4.68 | 95.38 | 4.46 |