Hugging Face Trending Papers

On-Demand Attention: Language Models Know When to Recall

Read the original on Hugging Face Trending Papers →

The paper introduces On‑Demand Attention (ODA), a local‑first decoding strategy that predicts when a pretrained language model would benefit from global attention. By training only a lightweight recall head, ODA selectively triggers global attention during generation, keeping pretrained weights unchanged and preserving the full key‑value cache for future recall. Experiments on Qwen, Gemma, and hybrid‑attention models show that ODA largely recovers performance lost with local attention while significantly cutting global reads, enabling faster long‑context inference.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at Hugging Face Trending Papers.

arXiv Computation and Language
Sep 18

On-Demand Attention: Language Models Know When to Recall

The paper introduces On‑Demand Attention (ODA), a decoding strategy that lets pretrained language models decide when to use global attention based on a lightweight recall head. ODA keeps the original model weights unchanged, only training the recall head, and can be implemented with GPU‑side conditional execution to reduce global reads. Experiments on Qwen, Gemma, and hybrid‑attention models show that ODA largely recovers performance lost by local attention while cutting the number of global attention operations.

By Haibo Feng, Ruiqi Liang, Hanyang Peng, Shiqi Yu
arXiv Machine Learning
3d ago

Memory Attention

The paper introduces Memory Attention (MA), a new attention mechanism for language models that replaces the traditional value projection with token-indexed memory combined with contextual keys. MA generates values by merging layer‑specific token memory with contextual information, allowing normalization to be folded into memory tables and reducing value construction to a simple lookup and addition. Experiments show that, with matched training token budgets and additional memory parameters, MA improves language modeling performance and average downstream task results across various attention configurations.

By Jiale Kang