Short-form Question Answering on Short-form QA Aggregate (Avg.) (test)
35.93EMVSPO + PRS_short
Evaluation Results
| Method | Links | |
|---|---|---|
| VSPO + PRS_shortBackbone=Qwen2.5-3B-Instruct2025.12 | 35.93 | |
| GRPO + PRS_shortBackbone=Qwen2.5-3B-Instruct2025.12 | 34.83 | |
| PPO + PRS_shortBackbone=Qwen2.5-3B-Instruct2025.12 | 33.33 | |
| SFTBackbone=Qwen2.5-3B-Instruct2025.12 | 17.6 | |
| UntrainedBackbone=Qwen2.5-3B-Instruct2025.12 | 9.6 |