Question Answering on 2WikiMultiHopQA 1,000 questions sample (val)
71.2Containment AccuracyFlowRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FlowRAGLLM=GPT-4o-mini, k=52026.06 | 71.2 | 65.2 | |
| LinearRAGLLM=GPT-4o-mini, k=52026.06 | 68.9 | 62.7 | |
| GFM-RAGLLM=GPT-4o-mini, k=52026.06 | 66.8 | 59.6 | |
| HippoRAGLLM=GPT-4o-mini, k=52026.06 | 66.1 | 59.9 | |
| HippoRAG2LLM=GPT-4o-mini, k=52026.06 | 62.7 | 55 | |
| LightRAGLLM=GPT-4o-mini, k=52026.06 | 55.2 | 39 | |
| E2GraphRAGLLM=GPT-4o-mini, k=52026.06 | 54.3 | 38.1 | |
| RAPTORLLM=GPT-4o-mini, k=52026.06 | 50.1 | 42.1 | |
| Vanilla RAGLLM=GPT-4o-mini, k=52026.06 | 48.6 | 43 | |
| G-retrieverLLM=GPT-4o-mini, k=52026.06 | 46.6 | 27.1 | |
| KGPLLM=GPT-4o-mini, k=52026.06 | 31.6 | 30 |