AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

vAttention: Verified Sparse Attention via Sampling

Aditya Desai, Kumar Krishna Agrawal, Shuo Yang, Alejandro Cuadron, Luis Gaspar Schroeder, Matei Zaharia, Joseph E. Gonzalez, Ion Stoica

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Block-sparse attention kernels

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 vAttention sampling to HashAttentionHashAttention · RULER32K-HARD (Llama-3.1-8B)average accuracy: 81.94 -> 86.56Improved
add vAttention sampling to HashAttentionHashAttention · RULER32K-HARD (DeepSeek-R1-Distill-Llama-8B)average accuracy: 60.70 -> 65.06Improved
add vAttention sampling to HashAttentionHashAttention · RULER32K-HARD (Mistral-7B)average accuracy: 54.66 -> 56.90Improved
add vAttention sampling to oracle-top-koracle-top-k · RULER32K-HARD (Llama-3.1-8B)average accuracy: 87.18 -> 88.61Improved
add vAttention sampling to oracle-top-koracle-top-k · RULER32K-HARD (DeepSeek-R1-Distill-Llama-8B)average accuracy: 64.87 -> 65.15No clear effect
add vAttention sampling to oracle-top-koracle-top-k · RULER32K-HARD (Mistral-7B)average accuracy: 64.37 -> 64.12No clear effect
add vAttention sampling to HashAttentionHashAttention · RULER32K easy datasets (Llama-3.1-8B)average accuracy: 85.80 -> 84.57Hurt
add vAttention sampling to HashAttentionHashAttention · RULER32K full benchmark (Llama-3.1-8B)average accuracy: 83.87 -> 85.57Improved
add vAttention sampling to oracle-top-koracle-top-k · LongBench (7 datasets, 10% sparsity)average score: 39.35 -> 41.33Improved
add vAttention sampling to HashAttentionHashAttention · LongBench (7 datasets, 10% sparsity)average score: 38.64 -> 40.04Improved
add vAttention to oracle-top-k for long generationoracle-top-k · AIME 2024 (Avg@4)accuracy: 36.66 -> 36.67No clear effect
use Hoeffding bound instead of CLT for sample budget (ablation)vAttention · RULER 16K, eps=0.1 delta=0.2samples per head: 2.8× more samples (average 874)Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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