Multi-hop Question Answering on TriviaQA fixed split (test)
66.41EMGraph-R1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Graph-R1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Graph-based2026.05 | 66.41 | 72.02 | |
| Graph-AdaptR1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Graph-based2026.05 | 64.06 | 72.77 | |
| Search-AdaptR1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 61.72 | 70.63 | |
| R1-SearcherBackbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 56.25 | 64.76 | |
| Search-R1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 51.56 | 61.03 | |
| R1Backbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=None2026.05 | 49.22 | 57.33 | |
| HippoRAG2Backbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 32.81 | 48.86 | |
| GraphRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 32.03 | 45.13 | |
| StandardRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Chunk-based2026.05 | 31.25 | 48.55 | |
| SFTBackbone=Qwen2.5-7B-Instruct, Interaction Type=Training, Knowledge Representation=Chunk-based2026.05 | 31.25 | 39.21 | |
| NaiveGenerationBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=None2026.05 | 28.91 | 47.73 | |
| HyperGraphRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 28.91 | 44.95 | |
| LightRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 25 | 40.18 | |
| PathRAGBackbone=GPT-4o-mini, Interaction Type=Prompt Engineering, Knowledge Representation=Graph-based2026.05 | 19.53 | 37.44 | |
| StandardRAGBackbone=Qwen2.5-7B-Instruct, Interaction Type=Prompt Engineering, Knowledge Representation=Chunk-based2026.05 | 8.59 | 24.9 | |
| NaiveGenerationBackbone=Qwen2.5-7B-Instruct, Interaction Type=Prompt Engineering, Knowledge Representation=None2026.05 | 7.03 | 24.51 |