AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

RESA: Bringing Back What Sparse Attention Ignores with Residual Estimation

Weihao Yang, Hao Huang, Ningke Li, Shihao Wang, Darong Yang, Yanqi Pan, Wen Xia, Shiyi Li, Xiangyu Zou

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Sparse attention for long-context LLMs

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 RESA residual estimationQuest · RULER MK3 (Llama-3.1-8B, 8K)accuracy: 24 -> 40Improved
add RESA residual estimationQuest · RULER MK3 (Llama-3.2-3B, 8K)accuracy: 3 -> 6Improved
add RESA residual estimationQuest · RULER MK3 (Mistral-7B, 8K)accuracy: 9 -> 27Improved
add RESA residual estimationIdeal-TopK · RULER MK3 (LWM-Text-7B)accuracy: 50 -> 70Improved
add RESA residual estimationIdeal-TopK · RULER MK3 (Mistral-7B)accuracy: 99 -> 99No clear effect
add RESA residual estimationArkVale · LongBench PR (Llama-3.1-8B)score: 89.19 -> 94.84Improved
add RESA residual estimationQuest · LongBench 2WMQA (Mistral-7B)score: 22.41 -> 30.63Improved
use mean of queries rather than last query as typical query (ablation)ArkVale · RULER MK3 (Llama3.2-3B)accuracy: 2 -> 6Improved
use mean of queries rather than last query as typical query (ablation)Quest · RULER MK3 (Llama3.2-3B)accuracy: 3 -> 5Improved
add RESA at 5% budgetQuest · RULER MK3 (Llama3.1-8B)accuracy: 50 -> 69Improved
increase prior weight lambda (ablation)RESA · RULER summary/comprehension tasksaccuracyMixed
add RESA to PSA/Ada-KV budgetingPSA · KV budget at <=2% accuracy lossKV budget: up to 33.2% budget reductionImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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