Question Answering on NQ-open v1.0 (test)
79.08A1ECC + MCH
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ECC + MCHBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=MCH2025.05 | 79.08 | 32.05 | 90.81 | 37.29 | 68.8 | |
| SE + MCHBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=MCH2025.05 | 78.47 | 30.13 | 89.73 | 42.57 | 69.66 | |
| PE + MCHBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=MCH2025.05 | 78.05 | 32.05 | 85.41 | 49.18 | 67.15 | |
| EigV + MCHBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=MCH2025.05 | 77.85 | 28.85 | 92.16 | 36.63 | 68.45 | |
| SEBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=None2025.05 | 76.4 | 32.69 | 90.54 | 37.62 | 65.88 | |
| SE + CTIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=CTI2025.05 | 74.12 | 31.41 | 87.57 | 42.57 | 65.13 | |
| ECC + CTIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=CTI2025.05 | 72.05 | 29.48 | 86.48 | 39.6 | 66.76 | |
| EigV + CTIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=CTI2025.05 | 71.22 | 24.36 | 87.3 | 37.95 | 65.06 | |
| PE + NLIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=NLI2025.05 | 70.81 | 42.31 | 80.54 | 52.81 | 66.5 | |
| SE + NLIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=NLI2025.05 | 70.39 | 42.95 | 87.03 | 49.51 | 66.92 | |
| ECC + NLIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=NLI2025.05 | 70.18 | 39.74 | 87.29 | 43.23 | 67.59 | |
| EigV + NLIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=NLI2025.05 | 69.98 | 38.46 | 88.65 | 41.91 | 67.29 | |
| PEBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=None2025.05 | 69.77 | 42.95 | 80.54 | 45.21 | 64.4 | |
| PE + CTIBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=CTI2025.05 | 67.91 | 40.39 | 78.6 | 48.85 | 65.66 | |
| EigVBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=None2025.05 | 54.66 | 14.1 | 85.4 | 27.72 | 63.03 | |
| ECCBackbone=Mistral-v0.3, Retriever=Contriver, Calibration=None2025.05 | 53 | 13.46 | 81.62 | 26.07 | 62.87 |