HG-RAG

HG-RAG is a retrieval-augmented generation model that uses graph-traversal over knowledge graphs, outperforming flat retrieval on hierarchical and multi-hop queries in environments with up to 800 nodes.

Two-hop subgraph: this entity, every entity it directly relates to, and every entity those neighbors relate to. Drag a node, scroll to zoom, click to inspect — or click any neighbor and re-center the atlas there.