Research project / HAT
Retrieval guided by known structure.
HAT uses session and document structure to retrieve stored vectors. On a synthetic 1,000-point benchmark it reported higher recall and faster construction than the tested HNSW configuration, while HNSW queries were faster.
The public record
Hierarchical Attention Tree: Extending LLM Context Through Structural Memory
- Scope
- Synthetic hierarchical retrieval: a 1,000-point, 128-dimensional comparison; a 500-5,000-point series; and a separate 60K-token retrieval example.
- Current limitation
- Known hierarchy favors HAT. HNSW queries were faster, downstream QA accuracy was not established, and larger-scale figure claims are not reconciled here.
- Available artifacts
- HAT repository and original paper PDF linked. The article separates reported experiments; exact run revisions, baseline tuning and uncertainty are not recorded here.
Editorial correction dated 2026-09-22; no new experiment or independent reproduction. The linked PDF is the original version and does not include these web corrections.