Question Answering on HotpotQA (EM, F1)
69.5EMReflective Confidence
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Reflective Confidence2025.12 | 69.5 | 82.3 | |
| Confidence-based Early-Stopping2025.12 | 66.8 | 80.1 | |
| Standard Prompting2025.12 | 65.2 | 78.5 |
| Method | Links | ||
|---|---|---|---|
| Reflective Confidence2025.12 | 69.5 | 82.3 | |
| Confidence-based Early-Stopping2025.12 | 66.8 | 80.1 | |
| Standard Prompting2025.12 | 65.2 | 78.5 |