Late Interaction outperforms dense models by 12 points in tests
tomaarsen · x · 2026-08-26
Tomaarsen compared Late Interaction against dense models. At matched training and backbones, the Late Interaction model won by +12 points over its Dense sibling. The strongest dense model, Qwen3-Embedding-4B, with 33x more active parameters, still fell short by 13 points.
Related event: Fine-tuning ColBERT on a single RTX 3090 beats general-purpose retrievers(14 posts)→
More from Research
- Intelligence is proportional to data: reorder tasks by data volume and the jagged frontier becomes a spiral — sebkrier · 2026-10-16
- Doc2Harness: New benchmark finds best coding agent passes only 53% across 329 cases — _akhaliq · 2026-10-16
- Stanford's Fei-Fei Li team open-sources OpenWAM, a unified world-action model framework for robots — jiqizhixin · 2026-10-16
- NVIDIA's TacitVLA: robots reason without reading their own thoughts, 3.4x lower latency — arankomatsuzaki · 2026-10-16
- Cohere Labs world model course session 4: DreamerV1-V4 architectures in one go — Cohere_Labs · 2026-10-16
- GPT-6.1 Sol audit finds 6% of arXiv papers had false main theorems, down to 1.5% recently — nitarshan · 2026-10-16