ResearchBenchmarksMulti-Hop Question Answering on HotpotQA (EM and F1)Follow58.2Exact Match (EM)DTKG (Full)41.87246.11150.3554.589Oct 18, 2025Evaluation ResultsMethodMethodLinksExact Match (EM)F1 ScoreDTKG (Full)Backbone=Llama 3.1:8BBackbone=Llama 3.1:8B2025.1058.271.5CoRAGBackbone=Llama 3.1:8BBackbone=Llama 3.1:8B2025.1056.369.8TOGBackbone=Llama 3.1:8BBackbone=Llama 3.1:8B2025.1050.264.5KGRBackbone=Llama 3.1:8BBackbone=Llama 3.1:8B2025.1048.562.4CRITICBackbone=Llama 3.1:8BBackbone=Llama 3.1:8B2025.1044.256.8COTBackbone=Llama 3.1:8BBackbone=Llama 3.1:8B2025.1042.555.4