How token pooling compresses multi-vector retrieval embeddings
tomaarsen · x · 2026-11-11
tomaarsen explains token pooling for multi-vector retrieval: models keep a vector per token so different query tokens match different document parts, but vectors add up. Pooling merges similar document vectors into cluster means via hierarchical clustering — the diagram shows a distance threshold τ cutting 10 token vectors down to 4 pooled vectors (2.5x fewer).
More from Research
- ArcGrav: a cross-model Agent Control Language to cut token overhead — ArchitecturalGravity · 2026-11-12
- Open-source dual-arm robot solves tangram puzzles with 90 demos, ~60% success — DanGrover · 2026-11-12
- FutureHouse's deslopper trains Kosmos to write indistinguishably from human experts — proteinrosh · 2026-11-12
- Fleuret amplifies debate: AI can't slash wet-lab experiments in unmapped biology — francoisfleuret · 2026-11-12
- Recursive's Eureka Machine ships Flint: a sparse 40B model with 2.5B active for local agentic coding — agihouse_org · 2026-11-12
- Recursive bets on open-endedness: agents running massive parallel experiments like scientist communities — agihouse_org · 2026-11-12