Glance Merges Reranker and Generator Into One Backbone for Slimmer RAG
airun-ai · hf · 2026-10-30
Glance from airun-ai unifies reranking and generation in a single backbone, replacing the separate pipeline of a reranker encoding k passages, selecting top-n, and a generator re-reading them.
- Lightweight adapters and pointer heads score all k candidate passages from one shared prefill of the context.
- After ranking, the model reuses the cached passage context to generate the answer, cutting the memory and compute overhead of re-encoding.
- The authors report improved retrieval and answer quality while reducing memory overhead.
More from Research
- Investor's Last Exam Benchmark: Top AI Models Ace Research but Fail at Investment Judgment — Huangfei Jiang · 2026-10-30
- SpecSlot Boosts Pixel-Space Image Generation: +2.84 FID on ImageNet-512 at Lower Cost — Jiongfan Yang · 2026-10-30
- Apple Trains UI Navigation Agents with Online RL Directly on Live iOS Apps — apple · 2026-10-30
- KAIST's L3G-World Generates Infinite Interactive Worlds in Real Time with Latent 3D Gaussian Memory — kaist-ai · 2026-10-30
- PhysInsert from Beihang and Meitu Achieves Physically Plausible Reference-Based Object Insertion — Yue Gong · 2026-10-30
- AB-Attention Speeds Up Video Diffusion Transformers with Training-Free Accurate Block-Sparse Attention — Bin Xu · 2026-10-30