Question Answering on QuALITY multiple choice (test)
86.09AccuracyGPT-4
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-4RAG Setting=Oracle RAG, Retriever=Perfect (Entire document in-context), Prompting=few-shot chain-of-thought2026.03 | 86.09 | 100 | |
| GPT-3.5RAG Setting=Oracle RAG, Retriever=Perfect (Entire document in-context), Prompting=few-shot chain-of-thought2026.03 | 72.6 | 100 | |
| EntiGraph CPTBase Model=Llama 3 8B, RAG Setting=Standard RAG, Prompting=few-shot chain-of-thought2026.03 | 62.6 | 99.63 | |
| Llama 3 8B BaseRAG Setting=Standard RAG, Prompting=few-shot chain-of-thought2026.03 | 60.35 | 99.63 |