Open-domain Question Answering on TriviaQA (TQA) few-shot Filtered (test)
77.3Exact Match (EM)Codex + REPLUG LSR
Evaluation Results
| Method | Links | |
|---|---|---|
| Codex + REPLUG LSRNumber of shots=16, Evaluation Protocol=In-context learning, Retrieval Corpus=Wikipedia (2018, December)2023.01 | 77.3 | |
| Codex + REPLUGNumber of shots=16, Evaluation Protocol=In-context learning, Retrieval Corpus=Wikipedia (2018, December)2023.01 | 76.8 | |
| Codex + ContrieverNumber of shots=16, Evaluation Protocol=In-context learning, Retrieval Corpus=MS-MARCO2023.01 | 76 | |
| AtlasNumber of shots=64, Evaluation Protocol=Finetuned2023.01 | 74.5 | |
| CodexNumber of shots=16, Evaluation Protocol=In-context learning2023.01 | 73.6 | |
| ChinchillaNumber of shots=64, Evaluation Protocol=In-context learning2023.01 | 64.6 |