Multi-Hop Question Answering on 2WikiMultiHopQA (test val)
43.9Exact Match (EM)DeepControl
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DeepControlBackbone Model=Qwen2.5-7B2026.02 | 43.9 | 47.9 | — | — | — | |
| DeepControlBackbone Model=Qwen2.5-3B2026.02 | 37.1 | 41.1 | — | — | — | |
| (ParaSet+SetCE)+CE5Base=Qwen3, Latency (sec/query)=0.2412026.07 | 35.3 | — | — | — | 38 | |
| (ParaSet+SetCE)+CE5Base=Contriever, Latency (sec/query)=0.1962026.07 | 35.1 | — | — | — | 39.2 | |
| ParaSet+SetCEBase=Qwen3, Latency (sec/query)=0.1642026.07 | 34.7 | — | — | — | 37.9 | |
| CE10Base=Qwen3, Latency (sec/query)=0.1142026.07 | 34.3 | — | — | — | 37 | |
| CE10Base=Contriever, Latency (sec/query)=0.0672026.07 | 31.5 | — | — | — | 35.2 | |
| CE5Base=Qwen3, Latency (sec/query)=0.1142026.07 | 31 | — | — | — | 33.5 | |
| SetCEBase=Qwen3, Latency (sec/query)=0.5852026.07 | 30.8 | — | — | — | 33.5 | |
| ListCEBase=Qwen3, Latency (sec/query)=0.5852026.07 | 30.7 | — | — | — | 33.5 | |
| ParaSet+SetCEBase=Contriever, Latency (sec/query)=0.1162026.07 | 28.7 | — | — | — | 33.2 | |
| BiBase=Qwen3, Latency (sec/query)=0.0732026.07 | 28.5 | — | — | — | 31.2 | |
| CE5Base=Contriever, Latency (sec/query)=0.0672026.07 | 28.4 | — | — | — | 30.9 | |
| ListCEBase=Contriever, Latency (sec/query)=0.5552026.07 | 25.1 | — | — | — | 27.7 | |
| SetCEBase=Contriever, Latency (sec/query)=0.5552026.07 | 24.1 | — | — | — | 26.7 | |
| BiBase=Contriever, Latency (sec/query)=0.0322026.07 | 21.9 | — | — | — | 24.3 | |
| E2GraphRAGCategory=Graph-based RAG2026.06 | — | — | 54.7 | 38.4 | — | |
| G-retrieverCategory=Graph-based RAG2026.06 | — | — | 47 | 26.9 | — | |
| GFM-RAGCategory=Graph-based RAG2026.06 | — | — | 67.3 | 59.1 | — | |
| HippoRAGCategory=Graph-based RAG2026.06 | — | — | 65.5 | 60.4 | — | |
| HippoRAG 2Category=Graph-based RAG2026.06 | — | — | 62.2 | 55.4 | — | |
| HyperGraphRAGCategory=Hypergraph-based RAG2026.06 | — | — | 68.8 | 64.1 | — | |
| HyperRAGCategory=Hypergraph-based RAG2026.06 | — | — | 66.3 | 69.1 | — | |
| HyperSUCategory=Hypergraph-based RAG2026.06 | — | — | 69.2 | 72.3 | — | |
| KGPCategory=Graph-based RAG2026.06 | — | — | 31.4 | 30.2 | — | |
| LightRAGCategory=Graph-based RAG2026.06 | — | — | 54.7 | 39.3 | — | |
| LinearRAGCategory=Graph-based RAG2026.06 | — | — | 69.6 | 64.2 | — | |
| MS-GraphRAGCategory=Graph-based RAG2026.06 | — | — | 48.1 | 29.6 | — | |
| RAPTORCategory=Graph-based RAG2026.06 | — | — | 50.5 | 41.7 | — |