RAM-Net: Linear-Time Sequence Modeling with Sparsely Addressable State

A scholar's reading copy on oatmeal paper: Georgia display type, pale sky-blue pencil marks, and a quiet chapter index pinned to the top.

muoncat.github.io/ramnet_webRAM-Net: Linear-Time Sequence Modeling with Sparsely Addressable State
Screenshot of https://muoncat.github.io/ramnet_web/, © the authors. Shown for reference. Request removal
Venue
NeurIPS 2026
Year
2026
First author
Kaicheng Xiao
Field
sequence modeling
Template
custom interactive paper companion (palette injected from data/palette.json, no template credit)
Theme
Light

Checked against the rendered page. Article width is limited at runtime by script; rendered boxes (965px article, 845px text) are reported instead of the CSS max-width.

Why it is here

A warm paper-beige palette with a serif display title, a horizontal in-page section nav and numbered sections. The five-panel comparison and the clean table add a scholarly look not found elsewhere in the picks, though the body text is small and low in contrast.

Overview

Color Palette

Typography

Georgia
RAM-Net: Linear-Time Sequence Modeling with Sparsely Addressable State Abstract · Method · Results · BibTeX — 0123456789
Segoe UI
RAM-Net: Linear-Time Sequence Modeling with Sparsely Addressable State Abstract · Method · Results · BibTeX — 0123456789

Spacing & Shape

Page Modules

Guidelines

More like this