AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

LimiX-2M: Mitigating Low-Rank Collapse and Attention Bottlenecks in Tabular Foundation Models

Yuanrui Wang, Xingxuan Zhang, Han Yu, Mingchao Hao, Gang Ren, Hao Yuan, Li Mao, Yunjia Zhang, Chun Yuan, Peng Cui

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
replace linear embedding with RaBEL (ablation)2M SNF baseline · first three layersnumerical rank: 58.41 -> 78.62Improved
replace linear embedding with RaBEL (ablation)2M SNF baseline · first three layersRank@99%: 13.94 -> 25.35Improved
replace linear embedding with RaBEL (ablation)2M SNF baseline · first three layersRank@95%: 6.73 -> 12.31Improved
replace MLP embedding with RaBEL (ablation)Transformer · BCCO-CLSAUC: 83.52 -> 85.04Improved
replace Periodic embedding with RaBEL (ablation)Transformer · BCCO-CLSAUC: 83.88 -> 85.04Improved
replace MLP embedding with RaBEL (ablation)Transformer · BCCO-REGRMSE: 0.4043 -> 0.3964Improved
replace Periodic embedding with RaBEL (ablation)Transformer · BCCO-REGR2: 0.6859 -> 0.7792Improved
replace MLP embedding with RaBEL (ablation)MLP · GC datasetMetric1: 0.7537 -> 0.9831Improved
replace MLP embedding with RaBEL (ablation)MLP · MA datasetMetric1: 0.8896 -> 0.8789Hurt
use orthogonal exponent-embedding init instead of Xavier or Kaiming (ablation)RaBEL · BCCO-CLSAUC (%): 89.03Improved
increase embedding dimension from 16 to 32 (ablation)RaBEL · BCCO-CLSAUC: 84.51 -> 85.17Improved
increase kernels from 16 to 64 (ablation)RaBEL · BCCO-CLSAUC: 84.20 -> 85.19Improved

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