AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Attribute Token Arithmetic: Disentangled and Continuous Semantic Control for Visual Autoregressive Models

Xindi Yang, Yicheng Wu, Cheng Zhang, Jianfei Cai, Tien-Tsin Wong

ECCV 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 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 raw text-token difference with learned disentangled directions (ablation)ATA (Infinity-2B) · GEditSemantic Consistency: 5.048 -> 5.374Improved
replace raw text-token difference with learned disentangled directions (ablation)ATA (Infinity-2B) · GEditPerceptual Quality: 7.447 -> 7.584Improved
replace raw text-token difference with learned disentangled directions (ablation)ATA (Infinity-2B) · GEditDisentanglement: 5.296 -> 5.455Improved
use ATA token directions instead of Concept Slider LoRA on same backbone (ablation)Infinity-2B · Control DatasetMean DeltaVQA: 0.532 -> 0.567Improved
use ATA token directions instead of Concept Slider LoRA on same backbone (ablation)Infinity-2B · Control DatasetMean I-LPIPS: 0.879 -> 0.895Improved
perturb prompt wording at test time (ablation)ATA (Infinity-2B) · Control DatasetMean DeltaVQA: 0.567 -> 0.541Hurt
perturb prompt wording at test time (ablation)ATA (Infinity-2B) · Control DatasetMean I-LPIPS: 0.895 -> 0.864Hurt
apply attribute direction to identity token instead of attribute token (ablation)ATA · qualitativeidentity preservationHurt
apply direction to unrelated token such as ground (ablation)ATA · qualitativeedit effectHurt
use ATA instead of VAREdit at extreme intensityATA · Control DatasetDeltaVQA: +6.95% -> +14.58%Improved
use ATA instead of Qwen-Image at moderate intensityATA · GEditSemantic Consistency: +6.20% -> +16.33%Improved
use ATA instead of VAREdit at moderate intensityATA · GEditPerceptual Quality: +5.75% -> +3.16%Hurt

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