Question Answering on NQ-Open (out-of-domain)
0.705PrecisionKP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| KPBackbone=Gemma-2-9B2026.02 | 0.705 | 0.602 | 0.649 | |
| VeriFY-NMBackbone=Gemma-2-9B2026.02 | 0.666 | 0.829 | 0.738 | |
| VeriFY-MMBackbone=Gemma-2-9B2026.02 | 0.643 | 0.869 | 0.739 | |
| SCBackbone=Gemma-2-9B2026.02 | 0.623 | 0.848 | 0.718 | |
| VeriFY-SMBackbone=Gemma-2-9B2026.02 | 0.622 | 0.888 | 0.732 | |
| KPBackbone=Gemma-2-2B2026.02 | 0.571 | 0.576 | 0.573 | |
| SCBackbone=Gemma-2-2B2026.02 | 0.538 | 0.717 | 0.615 | |
| BaseBackbone=Gemma-2-9B2026.02 | 0.535 | — | — | |
| VeriFY-NMBackbone=Gemma-2-2B2026.02 | 0.523 | 0.749 | 0.616 | |
| VeriFY-MMBackbone=Gemma-2-2B2026.02 | 0.496 | 0.808 | 0.615 | |
| VeriFY-SMBackbone=Gemma-2-2B2026.02 | 0.487 | 0.86 | 0.622 | |
| BaseBackbone=Gemma-2-2B2026.02 | 0.404 | — | — |