Conversation Question Answering on LOCOMO (test)
36.75RAG F1HippoRAG2 + DeepRefine-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| HippoRAG2 + DeepRefine-8BConstructor=AR1, Refinement Model=DeepRefine-8B2026.05 | 36.75 | |
| HippoRAG2 + DeepRefine-4BConstructor=AR1, Refinement Model=DeepRefine-4B2026.05 | 36.52 | |
| HippoRAG + DeepRefine-4BConstructor=AR1, Refinement Model=DeepRefine-4B2026.05 | 36.13 | |
| HippoRAG2Constructor=AR12026.05 | 35.75 | |
| HippoRAG + DeepRefine-8BConstructor=AR1, Refinement Model=DeepRefine-8B2026.05 | 35.64 | |
| HippoRAGConstructor=AR12026.05 | 35.02 | |
| HippoRAG + DeepRefine-4BConstructor=Naive, Refinement Model=DeepRefine-4B2026.05 | 33.73 | |
| HippoRAG + DeepRefine-8BConstructor=Naive, Refinement Model=DeepRefine-8B2026.05 | 33.56 | |
| HippoRAG2Constructor=Naive2026.05 | 33.33 | |
| HippoRAGConstructor=Naive2026.05 | 33.25 | |
| HippoRAG2 + DeepRefine-4BConstructor=Naive, Refinement Model=DeepRefine-4B2026.05 | 32.52 | |
| HippoRAG2 + DeepRefine-8BConstructor=Naive, Refinement Model=DeepRefine-8B2026.05 | 32.42 | |
| ToG + DeepRefine-4BConstructor=AR1, Refinement Model=DeepRefine-4B2026.05 | 27.84 | |
| Subgraph + DeepRefine-8BConstructor=AR1, Refinement Model=DeepRefine-8B2026.05 | 27.79 | |
| Subgraph + DeepRefine-4BConstructor=AR1, Refinement Model=DeepRefine-4B2026.05 | 27.71 | |
| Subgraph + DeepRefine-8BConstructor=Naive, Refinement Model=DeepRefine-8B2026.05 | 27.49 | |
| SubgraphConstructor=AR12026.05 | 26.65 | |
| Subgraph + DeepRefine-4BConstructor=Naive, Refinement Model=DeepRefine-4B2026.05 | 26.53 | |
| ToG + DeepRefine-8BConstructor=AR1, Refinement Model=DeepRefine-8B2026.05 | 26.16 | |
| SubgraphConstructor=Naive2026.05 | 25.86 | |
| ToG + DeepRefine-8BConstructor=Naive, Refinement Model=DeepRefine-8B2026.05 | 24.05 | |
| ToG + DeepRefine-4BConstructor=Naive, Refinement Model=DeepRefine-4B2026.05 | 23.81 | |
| ToGConstructor=AR12026.05 | 23.64 | |
| ToGConstructor=Naive2026.05 | 22.64 | |
| ToG + DeepRefine-8BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-8B2026.05 | 7.39 | |
| ToGConstructor=Graphify (LLM-Wiki)2026.05 | 7.3 | |
| ToG + DeepRefine-4BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-4B2026.05 | 7.08 | |
| Subgraph + DeepRefine-8BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-8B2026.05 | 5.95 | |
| Subgraph + DeepRefine-4BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-4B2026.05 | 5.75 | |
| SubgraphConstructor=Graphify (LLM-Wiki)2026.05 | 5.24 | |
| HippoRAG + DeepRefine-4BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-4B2026.05 | 4.5 | |
| HippoRAG + DeepRefine-8BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-8B2026.05 | 4.33 | |
| HippoRAG2 + DeepRefine-8BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-8B2026.05 | 4.09 | |
| HippoRAG2 + DeepRefine-4BConstructor=Graphify (LLM-Wiki), Refinement Model=DeepRefine-4B2026.05 | 3.69 | |
| HippoRAG2Constructor=Graphify (LLM-Wiki)2026.05 | 3.57 | |
| HippoRAGConstructor=Graphify (LLM-Wiki)2026.05 | 3.44 |