AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

A Theoretical Analysis of Mamba’s Training Dynamics: Filtering Relevant Features for Generalization in State Space Models

Mugunthan Shandirasegaran, Hongkang Li, Songyang Zhang, Meng Wang, Shuai Zhang

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Mamba architectures · 1 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 input-dependent gating (ablation)Mamba block · locality-structured synthetic datatest accuracyImproved
add input-dependent gating (ablation)Mamba block · majority-voting synthetic datatest accuracyImproved
increase majority-voting gap alpha_r - alpha_cMamba, GD · majority-voting synthetic dataepochs to convergeImproved
increase separation Delta L of class-relevant tokensMamba, GD · locality-structured synthetic dataconvergence speedHurt
vary feature dimension d among 32, 64, 128 (ablation)Mamba · majority-voting synthetic dataqualitative behavior: d in {32, 64, 128}No clear effect
vary confusion fraction alpha_c among 0.17, 0.20, 0.23 (ablation)Mamba · majority-voting synthetic dataagreement with theory: alpha_c 0.17, 0.20, 0.23No clear effect
use Mamba instead of global attentionsingle-layer model · locality-structured synthetic dataaccuracyImproved

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