AI Paper Atlas

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

Hidden Dangers of Compositional Generation: Diagnosing Semantic Safety Failures in Text-to-Image Models

Haoming Yang, Ke Ma, Ligong Zhang, Xiaojun Jia, Yingfei Sun, Qianqian Xu, Qingming Huang

CVPR 2026 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters

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 fine-grained decomposition and reassembly on top of visual metaphors (ablation)CoRA · DALL-E 3ASR: 0.081 -> 0.644Improved
add decomposition and reassembly on top of visual metaphors (ablation)CoRA · Cogview4ASR: 0.444 -> 0.733Improved
add decomposition and reassembly on top of visual metaphors (ablation)CoRA · Cogview4Prompt perplexity: 97.193 -> 37.278Improved
add decomposition and reassembly on top of visual metaphors (ablation)CoRA · DALL-E 3Inception Score: 1.000 -> 3.813Improved
use stronger auxiliary model Qwen3-235B instead of Qwen3-8B (ablation)CoRA · five T2I modelsASR: 0.73 -> 0.74 (Cogview4)No clear effect
use weaker auxiliary model Qwen2-7B instead of Qwen3-8B (ablation)CoRA · SDXLASR: 0.59 -> 0.57No clear effect
use CoRA instead of DACAT2I safety attack · DALL-E 3ASR: 0.407 -> 0.644Improved
use CoRA instead of DACAT2I safety attack · GPT-4o web interfaceASR: 0.433 -> 0.667Improved
use CoRA instead of MMA-Diffusion on DALL-E 3 semantic consistencyT2I safety attack · DALL-E 3Semantic Consistency: 0.261 -> 0.258No clear effect
use CoRA instead of Ring-a-Bell on prompt perplexityT2I safety attack · Cogview4Prompt perplexity: 12752.81 -> 37.28Improved

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