Open-domain Question Answering on TriviaQA (TQA) (dev)
72.7EMPAQ + FiE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PAQ + FiEModel Scale=Large2022.11 | 72.7 | — | |
| RFiD-large#para=990M, implementation=our implementation2023.05 | 72.7 | — | |
| FiEModel Scale=Large2022.11 | 71.6 | — | |
| FiD-large#para=990M, implementation=our implementation2023.05 | 71.6 | — | |
| RFiD-base#para=440M, implementation=our implementation2023.05 | 69.6 | — | |
| PAQ + FiEModel Scale=Base2022.11 | 69.5 | — | |
| FiD-base#para=440M, implementation=our implementation2023.05 | 68.6 | — | |
| FiEModel Scale=Base2022.11 | 68.3 | — | |
| Entities as Experts#Params=360M, #Encoded=12021.10 | — | 43.2 | |
| Fusion-in-Decoder#Params=440M, #Encoded=1002021.10 | — | 65 | |
| RAG#Params=620M, #Encoded=1002021.10 | — | 56.8 | |
| REALM#Params=330M, #Encoded=52021.10 | — | 55.8 | |
| T5-11B#Params=11B, #Encoded=12021.10 | — | 42.3 | |
| TOME-1#Params=220M, #Encoded=12021.10 | — | 50.8 | |
| TOME-2#Params=220M, #Encoded=12021.10 | — | 54.6 |