Knowledge Updating on Context-Derived QA Generalization (test)
68AccuracyPASTA
Evaluation Results
| Method | Links | |
|---|---|---|
| PASTAIteration=5, Stage=Final2026.06 | 68 | |
| PASTAIteration=6, Stage=Final2026.06 | 67 | |
| PASTAIteration=4, Stage=Final2026.06 | 66 | |
| PASTAIteration=2, Stage=Final2026.06 | 65 | |
| PASTAIteration=3, Stage=Final2026.06 | 64 | |
| PASTAIteration=7, Stage=Final2026.06 | 64 | |
| PASTAIteration=1, Stage=Final2026.06 | 54 | |
| Baseline (SFT only)Iteration=7, Stage=Final2026.06 | 26 | |
| Baseline (SFT only)Iteration=5, Stage=Final2026.06 | 19 | |
| Baseline (SFT only)Iteration=6, Stage=Final2026.06 | 18 | |
| Baseline (SFT only)Iteration=2, Stage=Final2026.06 | 16 | |
| Baseline (SFT only)Iteration=3, Stage=Final2026.06 | 13 | |
| Baseline (SFT only)Iteration=4, Stage=Final2026.06 | 9 | |
| Baseline (SFT only)Iteration=1, Stage=Final2026.06 | 7 |