Question Answering on QuALITY hard
76.2AccuracyRAPTOR + GPT-4
Evaluation results
| Method | Links | |
|---|---|---|
| RAPTOR + GPT-4Language Model=GPT-42024.01 | 76.2 | |
| CoLISA (DeBERTaV3-large)Reader=DeBERTaV3-large2024.01 | 54.7 | |
| DPR and DeBERTaV3-largeRetriever=DPR, Reader=DeBERTaV3-large2024.01 | 46.1 | |
| Longformer-baseVariant=base2024.01 | 35.3 |