Question Answering on Multi-Dataset Average (NQ, TriviaQA, HotpotQA, StrategyQA, 2WikiMHQA)
55Average ScoreRA-ISF
Evaluation results
| Method | Links | |
|---|---|---|
| RA-ISFModel=GPT-3.5, Retrieval Configuration=With Retrieval2024.03 | 55 | |
| SKRknnModel=GPT-3.5, Retrieval Configuration=With Retrieval2024.03 | 47.6 | |
| IRCoTModel=GPT-3.5, Retrieval Configuration=With Retrieval2024.03 | 46.5 | |
| Least-to-mostModel=GPT-3.5, Retrieval Configuration=Without Retrieval2024.03 | 46.3 | |
| RA-ISFModel=Llama-2-13b, Retrieval Configuration=With Retrieval2024.03 | 46 | |
| RAGModel=GPT-3.5, Retrieval Configuration=With Retrieval2024.03 | 44.2 | |
| Self-RAG13BModel=Llama-2-13b, Retrieval Configuration=With Retrieval2024.03 | 44.1 | |
| DirectModel=GPT-3.5, Retrieval Configuration=Without Retrieval2024.03 | 41.8 | |
| REPLUGModel=Llama-2-13b, Retrieval Configuration=With Retrieval2024.03 | 38.6 | |
| SKRknnModel=Llama-2-13b, Retrieval Configuration=With Retrieval2024.03 | 36 | |
| IRCoTModel=Llama-2-13b, Retrieval Configuration=With Retrieval2024.03 | 34 | |
| RAGModel=Llama-2-13b, Retrieval Configuration=With Retrieval2024.03 | 33.9 | |
| Least-to-mostModel=Llama-2-13b, Retrieval Configuration=Without Retrieval2024.03 | 32.9 | |
| Vanilla LMModel=Llama-2-13b, Retrieval Configuration=Without Retrieval2024.03 | 27.1 |