AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Mixture of States: Routing Token-Level Dynamics for Multimodal Generation

Haozhe Liu, Ding Liu, Mingchen Zhuge, Zijian Zhou, Tian Xie, Sen He, Yukang Yang, Shuming Liu, Yuren Cong, Jiadong Guo, Hongyu Xu, Ke Xu and 7 more

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 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
feed latent and timestep to the router in addition to the prompt (ablation)MoS router · MJHQFID: 21.12 -> 20.15Improved
add noised latent to router input (without timestep) (ablation)MoS router · MJHQFID: 21.12 -> 21.89Hurt
add latent and timestep to router input (ablation)MoS router · MJHQCLIP score: 21.40 -> 21.74Improved
use token-specific routing instead of sample-wise routing (ablation)MoS router · MJHQ ablationFID and CLIPImproved
use MoS adaptive routing instead of fixed-layer cross-attention (ablation)MoS vs cross-attention · GenEvaloverall: 0.74 -> 0.79Improved
use MoS adaptive routing instead of fixed-layer cross-attention (ablation)MoS vs cross-attention · DPG-Benchoverall: 83.40 -> 85.61Improved
use MoS routing instead of MoT one-to-one layer correspondence (ablation)MoS vs MoT (PLM-1B towers) · GenEval and DPGscoreImproved
add self-CoT reasoning captionMoS-L · WISEoverall: 0.54 -> 0.65Improved
add self-CoT reasoning captionMoS-S · WISEoverall: 0.47 -> 0.55Improved
use top-k=2 instead of k=1 or larger k (ablation)MoS router · ablation (training curves)performanceImproved
add epsilon-greedy exploration (ablation)MoS router · training convergenceconvergence speedImproved
use separate RMSNorm per modality in the router (ablation)MoS router · MJHQ FID/CLIPFID and CLIPImproved

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