List your company

Long narrative understanding QA on NarrativeQA (Acc %)

55AccuracyHGMEM
32.1238.064449.94Dec 30, 2025

Evaluation results

MethodLinks
HGMEMBase LLM=GPT-4o, Type=Ours, Working Memory=true2025.1255
ComoRAGBase LLM=GPT-4o, Type=Multi-step RAG, Working Memory=true2025.1254
GraphRAGBase LLM=GPT-4o, Type=Traditional RAG, Working Memory=false2025.1253
NaiveRAGBase LLM=GPT-4o, Type=Traditional RAG, Working Memory=false2025.1252
HGMEMBase LLM=Qwen2.5-32B-Instruct, Type=Ours, Working Memory=true2025.1251
DeepRAGBase LLM=GPT-4o, Type=Multi-step RAG, Working Memory=true2025.1245
LightRAGBase LLM=GPT-4o, Type=Traditional RAG, Working Memory=false2025.1244
GraphRAGBase LLM=Qwen2.5-32B-Instruct, Type=Traditional RAG, Working Memory=false2025.1244
DeepRAGBase LLM=Qwen2.5-32B-Instruct, Type=Multi-step RAG, Working Memory=true2025.1244
ComoRAGBase LLM=Qwen2.5-32B-Instruct, Type=Multi-step RAG, Working Memory=true2025.1244
LightRAGBase LLM=Qwen2.5-32B-Instruct, Type=Traditional RAG, Working Memory=false2025.1240
NaiveRAGBase LLM=Qwen2.5-32B-Instruct, Type=Traditional RAG, Working Memory=false2025.1237
HippoRAG v2Base LLM=GPT-4o, Type=Traditional RAG, Working Memory=false2025.1234
HippoRAG v2Base LLM=Qwen2.5-32B-Instruct, Type=Traditional RAG, Working Memory=false2025.1233