Knowledge Graph RAG Boosts Multi-Hop Retrieval Accuracy by 54%

Vector Institute released an Entity-Based Knowledge Graph RAG architecture that tackles multi-hop retrieval failures in complex enterprise documents. By concatenating entity names with source document titles before embedding, it lifted GPT-4o-mini's multi-hop accuracy on SEC 10-Q data from 36% to 56%.

2026-10-09 ~ 2026-10-09 · 3 related posts

• Vector Institute shows hyphenating entities with doc titles fixes RAG's lost-context problem — VectorInst · 2026-10-09
• Vector Institute's Knowledge Graph RAG Boosts Multi-Hop Retrieval Accuracy by Up to 54% — VectorInst · 2026-10-09
• Knowledge graphs lift GPT-4o-mini multi-hop retrieval from 36% to 56% at 0.05s cost, Vector Institute reports — VectorInst · 2026-10-09