General Question Answering on ExpertQA
0.2385RewardDAPO+START
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DAPO+STARTConfiguration=DAPO+START2026.03 | 0.2385 | 65.29 | |
| DAPOConfiguration=DAPO2026.03 | 0.2352 | — | |
| GRPO+STARTEvaluation Protocol=Post-trained2026.03 | 0.2201 | 59.24 | |
| GRPOEvaluation Protocol=Post-trained2026.03 | 0.2182 | — | |
| GRPO-MAEvaluation Protocol=Post-trained2026.03 | 0.214 | 34.08 | |
| GRPO+STARTEvaluation Protocol=T_post + A_pre2026.03 | 0.1876 | 68.15 | |
| GRPO-MAEvaluation Protocol=T_post + A_pre2026.03 | 0.1859 | 54.14 | |
| GRPOEvaluation Protocol=T_post + A_pre2026.03 | 0.1849 | — |