Structural Knowledge Graph Evaluation on CogEval-Bench averages over 6 scenarios
0.476HarmonyCOGNIFOLD
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| COGNIFOLDLLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0.476 | 35.8 | 0.733 | 0.361 | 0.833 | 0.327 | 0.546 | 0.624 | 4.6 | 0.274 | 0.614 | |
| GraphRAGLLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0.323 | 23.2 | 0.541 | 0 | 0.833 | 0.002 | 0.873 | 0.968 | 1.2 | 0.317 | 0 | |
| OpenIE KGLLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0.138 | 8.1 | 0.492 | 0 | 0.75 | 0.005 | 0.848 | 0.968 | 0.3 | 0.323 | 0 | |
| ZepLLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0.138 | 8.1 | 0.494 | 0 | 0.917 | 0.704 | 0.784 | 0.507 | 0.5 | 0.242 | 0 | |
| HippoRAG 2LLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0.095 | 5.8 | 0.341 | 0 | 0.917 | 0.716 | 0.759 | 0 | 0.2 | 0.31 | 0 | |
| CogneeLLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0.094 | 5.3 | 0.454 | 0 | 1 | 0.2 | 0.714 | 0.54 | 0.5 | 0.352 | 0 | |
| Mem0LLM=gpt-4o-mini, Embedding Model=text-embedding-3-small, Ingestion Mode=stream ingestion2026.05 | 0 | 0 | 0 | 0 | 1 | 0.382 | 0.434 | 0 | 1 | 0.189 | 0 |