AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

LoKiFormer: Locality-aware Attention with Decoupled Knowledge Memory for Efficient Large Language Model Pretraining

Qiuwu Chen, Zimo Liu, Yuchen Li, Ying Sun, Yifan Zhang, Zhijie Qiu, Zeng You, Ryan Dong, Simeng Ma, Yaofo Chen, Mingkui Tan

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Long-context KV and memory attention · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Local Fusion Attention (LFA) (ablation)baseline 5B · Training PPL at 10k stepsperplexity: 31.82 -> 30.57Improved
add Knowledge Memory Module (KMM) on top of LFA (ablation)baseline+LFA 5B · Training PPL at 10k stepsperplexity: 30.57 -> 28.50Improved
add LFA (ablation)baseline 5B · MMLU zero-shotaccuracy: 17.9 -> 21.2Improved
add KMM (ablation)baseline 5B · MMLU zero-shotaccuracy: 17.9 -> 22.9Improved
add both LFA and KMM (ablation)baseline 5B · MMLU zero-shotaccuracy: 17.9 -> 25.7Improved
increase LFA kernel size from 2 to 4 (ablation)LFA 5B · Training PPLperplexity: 31.06 -> 30.59Improved
use head-aligned groups g=h instead of g=1 in LFA (ablation)LFA 5B · Training PPL at 10k stepsperplexity: 31.07 -> 30.49Improved
use per-dimension groups g=d instead of g=1 in LFA (ablation)LFA 5B · Training PPL at 10k stepsperplexity: 31.07 -> 31.49Hurt
increase knowledge fields F from 0 to 32 (ablation)KMM 5B · Training PPL cc-enperplexity: 31.62 -> 30.78Improved
increase knowledge fields F from 32 to 64 (ablation)KMM 5B · Training PPL cc-enperplexity: 30.78 -> 30.35Improved
increase knowledge fields F from 64 to 128 (ablation)KMM 5B · Training PPL cc-enperplexity: 30.35 -> 29.88Improved
add LFA and KMM to the baseline (ablation)LoKiFormer-7B · Training throughput (4K context)tokens/second: 975 -> 950 tokens/secondNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy