Signals 4 · free daily AI digest

Learning Length-Extrapolatable Recurrent Models

Paper recorded by Signals 4 on 2026-09-08 in cs.LG. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-08 on arXiv · recorded by Signals 4 on 2026-09-09

Category: cs.LG · 机器学习 · first seen 2026-09-09

Abstract

Recurrent models provide a natural path to long-context modeling, yet models trained with backpropagation through time (BPTT) often fail beyond their training horizon. Classical analyses emphasize gradients that vanish or explode along temporal paths. However, dense per-token losses can still train a shared recurrent rule despite severe decay, showing that decay alone does not determine whether le

Read on arXiv →

#94 most recent of 215 cs.LG papers we have recorded · ↑ newer: TimeCues Studio: A Workspace for Music Annotation and Algorithm Protot · ↓ older: Silver Rate Is (Almost) Optimal for Gradient Descent Acceleration
Cite this page: Learning Length-Extrapolatable Recurrent Models: the #94 most recent of 215 cs.LG papers we have recorded (as of 2026-09-08). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/learning-length-extrapolatable-recurrent-models.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.LG papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions