AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Gated Condition Injection without Multimodal Attention: Towards Controllable Linear-Attention Transformers

Yuhe Liu, Zhenxiong Tan, Yujia Hu, Songhua Liu, Xinchao Wang

CVPR 2026 · Image generation › Autoregressive and latent generation · approach: High-resolution diffusion synthesis · 1 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
add token-wise gated condition injection (ablation)OminiControl-style SANA · COCO val Canny-to-imageFID: 22.6 -> 19.0Improved
add gated condition injection vs OminiControlSANA · COCO val ColorizationFID: 24.95 -> 10.28Improved
add gated condition injection vs ControlNetSANA · COCO val HED-to-imageMSE: 2320 -> 1168Improved
add gated condition injection vs OminiControlSANA · COCO val DeblurringMSE: 120 -> 14Improved
add gated condition injection vs OminiControlSANA · COCO val Canny-to-image (512)F1: 0.23 -> 0.26Improved
add gated condition injection vs OminiControlSANA · COCO val Depth-to-image (512)MSE: 803 -> 626Improved
add attention interaction between latent and condition tokens (ablation)GateControl · COCO val Canny-to-imageFID: 20.0 -> 19.0Improved
use token-wise instead of element-wise gating (ablation)GateControl · COCO val Canny-to-imageFID: 18.8 vs 19.0No clear effect
place gate after self-attention instead of after Mix-FFN (ablation)GateControl · COCO val Canny-to-imageFID: 18.2 vs 19.0No clear effect
compute gate scores from features before self-attention instead of after (ablation)GateControl · COCO val Canny-to-imageFID: 20.3 -> 19.0Improved
use gating instead of direct summation of condition features (ablation)GateControl · Canny-to-imagetraining lossImproved
place gate after self-attention instead of after cross-attention (ablation)GateControl · Canny-to-imagetraining lossImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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