Question Answering on TruthfulQA Semantic-level split o=3
98.1Accuracy
W/O Decontamination
Updated 4d ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.01 | 98.1 | 0.514 | |
2026.01 | 97.1 | 0.722 | |
2026.01 | 97.1 | 0.722 | |
2026.01 | 95.7 | 0.49 | |
2026.01 | 95.2 | 0.421 | |
2026.01 | 95.1 | 0.485 | |
2026.01 | 94.7 | 0.416 | |
2026.01 | 91.4 | 0.665 | |
2026.01 | 90 | 0.369 | |
2026.01 | 85.2 | 0.385 | |
2026.01 | 63.4 | 0.385 | |
2026.01 | 63.2 | 0.101 | |
2026.01 | 62.2 | 0.091 | |
2026.01 | 60.8 | 0.141 | |
2026.01 | 57.9 | 0.33 | |
2026.01 | 55.1 | 0.302 | |
2026.01 | 53.1 | - | |
2026.01 | 52.6 | 0.005 | |
2026.01 | 51.9 | 0.053 | |
2026.01 | 46.7 | - | |
2026.01 | 24.9 | - |