Multi-hop Question Answering on Multi-hop QA Datasets calculated average (test)
51.43EMGraph-AdaptR1
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Graph-AdaptR1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Graph-based2026.05 | 51.43 | 60.74 | 61.88 | |
| Graph-R1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Graph-based2026.05 | 50.65 | 59.27 | 60.46 | |
| Search-AdaptR1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 42.58 | 50.68 | 65.13 | |
| Search-R1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 38.29 | 43.8 | 53.06 | |
| R1-SearcherBackbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 34.51 | 42.29 | 51.26 | |
| R1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=None2026.05 | 25.91 | 33.12 | — | |
| StandardRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Chunk-based2026.05 | 18.1 | 32.05 | 52.68 | |
| SFTBackbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 15.57 | 24.01 | — | |
| HippoRAG2Backbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 13.8 | 25.82 | 36.41 | |
| HyperGraphRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 13.15 | 29.4 | 61.82 | |
| GraphRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 12.5 | 24.87 | 32.48 | |
| NaiveGenerationBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=None2026.05 | 11.36 | 25.92 | — | |
| LightRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 9.77 | 24.24 | 47.42 | |
| PathRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 7.03 | 20.02 | 46.71 | |
| StandardRAGBackbone=Qwen2.5-7B-Instruct, Interaction Type=Prompt Engineering, Knowledge Representation=Chunk-based2026.05 | 5.34 | 15.39 | 52.67 | |
| NaiveGenerationBackbone=Qwen2.5-7B-Instruct, Interaction Type=Prompt Engineering, Knowledge Representation=None2026.05 | 3.12 | 14.2 | — |