Short-form Question Answering on PopQA (test)
36.34Exact MatchVSPO + PRS_short
Evaluation Results
| Method | Links | |
|---|---|---|
| VSPO + PRS_shortBackbone=Qwen2.5-3B-Instruct2025.12 | 36.34 | |
| GRPO + PRS_shortBackbone=Qwen2.5-3B-Instruct2025.12 | 35.62 | |
| PPO + PRS_shortBackbone=Qwen2.5-3B-Instruct2025.12 | 35.62 | |
| SFTBackbone=Qwen2.5-3B-Instruct2025.12 | 10.4 | |
| UntrainedBackbone=Qwen2.5-3B-Instruct2025.12 | 8.49 |