Long-Context Question Answering on SQuAD 2K
84.9Answer F1 ScoreIN-CONTEXT
Evaluation Results
| Method | Links | |
|---|---|---|
| IN-CONTEXT2026.02 | 84.9 | |
| GEN ADAPTER2026.02 | 39.9 | |
| SHINETraining Data Scale=Full, Training Data Source=MQA + 1QA2026.02 | 37.5 | |
| SHINETraining Data Scale=1/2, Training Data Source=1QA2026.02 | 34.6 | |
| SHINETraining Data Scale=1/2, Training Data Source=MQA2026.02 | 33.8 | |
| NAIVE2026.02 | 22 |