Hierarchical token pooling halves multi-vector indexes at 100.6% of BEIR performance
tomaarsen · x · 2026-08-18
The pain point of multi-vector retrieval is index size: 4,874 NQ passages become 608,414 token vectors — 311.5 MB vs 20 MB for a 1024d dense index, about 16x. HierarchicalTokenPooling (Clavié, Chaffin & Adams) clusters each document's token vectors with Ward linkage and keeps 1/poolfactor of them; at poolfactor=2 the index halves while retaining 100.6% of unpooled BEIR performance, usable per call, standalone, or baked into the model. The other escape routes: a real late-interaction index, or use it as a reranker.
More from coding & agent
- ai& becomes an official built-in Cline provider with 13 open-weight models — DavidBennett__ · 2026-10-16
- x402 as a free GTM channel: agents discover 100+ tools, no API keys needed — 0xJeff · 2026-10-16
- Sam Altman: stop getting better at prompts, build loops and graphs that prompt for you — goyalshaliniuk · 2026-10-16
- POV ski game on real Matterhorn elevation data, built with Claude Opus 5.5 — techartist_ · 2026-10-16
- Poisoned SERPs broke rank tracking — a Claude MCP + Notion workaround — VeryWellVersed · 2026-10-16
- OpenAI's Live API adds Decisions API for ~200ms real-time voice app decisions — juberti · 2026-10-16