AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regression

Yifei Zuo, Yutong Yin, Zhichen Zeng, Ang Li, Banghua Zhu, Zhaoran Wang

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Linear attention approximations

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
use local linear attention instead of softmax attention (local constant)single attention layer, test-time only · piecewise-linear test-time regression (d=64, L=1024)position-wise MSEImproved
use LLA instead of MesaNetsingle attention layer, test-time only · piecewise-linear test-time regression, first segmentposition-wise MSEHurt
add more in-segment data (ablation)LLA vs softmax attention · piecewise-linear test-time regressionposition-wise MSEMixed
increase data dimension d (ablation)LLA advantage over baselines · piecewise-linear test-time regressionMSE ratio vs LLAImproved
use LLA instead of Softmax, Mamba, GLA, Hyena, Gated DeltaNettwo-layer models without MLPs · in-context regression on non-stationary data (dx=dy=32)test errorImproved
use LLA instead of softmax and linear baselinesd=128, 2 heads, no short convolution · MQAR associative recallrecall accuracyImproved
use LLA instead of softmax attentionpermutation state tracking · permutation state-trackingtest accuracyNo clear effect
use blockwise FlashLLA Triton kernel instead of naive implementationLLA · profiling on H200, d=128memoryImproved

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