Multi-hop Question Answering on HotpotQA (SubEM, LLM-Acc)
65.7SubEMEHRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EHRAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 65.7 | 69.3 | |
| LinearRAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 64.3 | 66.5 | |
| HippoRAG2Generation Strategy=Retrieval-Augmented-Generation Methods2026.04 | 62.9 | 64.3 | |
| GFM-RAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 62.7 | 65.6 | |
| KGPGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 61.5 | 60.9 | |
| E2GraphRAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 61 | 63.9 | |
| LightRAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 60.3 | 59.5 | |
| GraphRAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 58.6 | 59.8 | |
| HippoRAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 57 | 59.3 | |
| RAPTORGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 55.9 | 58.3 | |
| Vanilla RAGGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 55.7 | 58.6 | |
| G-retrieverGeneration Strategy=Retrieval-Augmented-Generation Methods2026.04 | 42.2 | 40.6 | |
| GPT-4o-miniGeneration Strategy=Direct Zero-shot LLM Inference2026.04 | 38.9 | 40.2 | |
| GPT-3.5-turboGeneration Strategy=Direct Zero-shot LLM Inference2026.04 | 33.4 | 43.2 | |
| llama-8BGeneration Strategy=Direct Zero-shot LLM Inference2026.04 | 31.1 | 27.3 | |
| Qwen3-8BGeneration Strategy=Direct Zero-shot LLM Inference2026.04 | 25.1 | 34.7 | |
| llama-13BGeneration Strategy=Direct Zero-shot LLM Inference2026.04 | 24.2 | 16.8 |