Clarifying Questions on SciQA (test)
26AccuracySwift
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SwiftMode=Thinking Mode2025.06 | 26 | 62 | 4.63 | 5.77 | 5.39 | 4.44 | 5.07 | 4.96 | |
| RefitMode=Thinking Mode2025.06 | 24 | 68 | 5.03 | 6.21 | 5.81 | 4.75 | 5.42 | 5.33 | |
| STaR-GATE-DMode=Thinking Mode2025.06 | 14 | 16 | 4.01 | 4.95 | 4.76 | 3.67 | 4.3 | 4.65 | |
| DPOMode=Thinking Mode2025.06 | 8 | 24 | 4.44 | 5.53 | 5.17 | 4.19 | 4.86 | 4.86 | |
| StepDPOMode=Thinking Mode2025.06 | 8 | 34 | 4.7 | 5.85 | 5.36 | 4.44 | 5.09 | 5.13 | |
| BaseMode=Thinking Mode2025.06 | 6 | 86 | 5.47 | 6.69 | 6.14 | 5.1 | 5.95 | 5.87 |