AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Revelio: Interpreting and leveraging semantic information in diffusion models

Dahye Kim, Xavier Thomas, Deepti Ghadiyaram

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Diffusion models as perception tools · 17 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
use up_ft1 decoder features instead of bottleneck features (ablation)Diff-C · Oxford-IIIT Pet (SD 1.5, t=25)top-1 accuracy: 69.97 -> 88.61Improved
use up_ft1 for coarse-grained datasets instead of bottleneck (label purity) (ablation)k-SAE · Caltech-101sigma_label: 9.35 -> 21.33Hurt
extract features at t=25 rather than t=200 (ablation)Diff-C up_ft1 · Oxford-IIIT Pet (SD 1.5)top-1 accuracy: 86.89 -> 88.61Improved
use t=200 features rather than t=25 for coarse dataset (ablation)k-SAE bottleneck · Caltech-101sigma_label: 9.35 -> 8.17Improved
use SD 1.5 instead of SD 2.1 (ablation)Diff-C up_ft1 · Oxford-IIIT Pet (t=25)top-1 accuracy: 84.74 -> 88.61Improved
add CLIP-inferred class prompt as text conditioning (ablation)Diff-C · Oxford-IIIT Pettop-1 accuracy: 88.69 -> 90.97Improved
add CLIP-inferred class prompt as text conditioning (ablation)Diff-C · FGVC-Aircrafttop-1 accuracy: 65.07 -> 64.98No clear effect
use a generic prompt 'A photo of a pet' instead of empty prompt (ablation)Diff-C up_ft1 · Oxford-IIIT Pettop-1 accuracy: -3.14Hurt
interleave diffusion up_ft1 features (t=25) with CLIP featuresLLaVA-Lightning · LLaVA-Bench (In-the-Wild)relative score: 56.6 -> 59.9Improved
interleave DINO-v2 features with CLIPLLaVA-Lightning · LLaVA-Bench (In-the-Wild)relative score: 56.6 -> 47.0Hurt
add diffusion features (t=25) to CLIP in LLaVALLaVA-Lightning · MM-Vetcorrectness score: +1.1Improved
use features from pixel-space DeepFloyd-IF at up_ft0 instead of SD 1.5Diff-C · Oxford-IIIT Pet and Caltech-101 (t=25)top-1 accuracy: +1.74% Pet, +2.85% Caltech-101Improved

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