Cover Image

Text-to-Image Diffusion Models Cannot Count, and Prompt Refinement Cannot Help

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Guo, Xuyang, Huo, Jiayan, Liang, Yingyu, Shi, Zhenmei, Song, Zhao, Zhang, Jiahao, Zhuang, Zhen
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Can You Count to Nine? A Human Evaluation Benchmark for Counting Limits in Modern Text-to-Video Models
by: Guo, Xuyang, et al.
Published: (2025)

T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video Generation
by: Guo, Xuyang, et al.
Published: (2025)

Your Vision-Language Model Can't Even Count to 20: Exposing the Failures of VLMs in Compositional Counting
by: Guo, Xuyang, et al.
Published: (2025)

T2VTextBench: A Human Evaluation Benchmark for Textual Control in Video Generation Models
by: Guo, Xuyang, et al.
Published: (2025)

T2VWorldBench: A Benchmark for Evaluating World Knowledge in Text-to-Video Generation
by: Chen, Yubin, et al.
Published: (2025)

Unraveling the Smoothness Properties of Diffusion Models: A Gaussian Mixture Perspective
by: Liang, Yingyu, et al.
Published: (2024)

RichSpace: Enriching Text-to-Video Prompt Space via Text Embedding Interpolation
by: Cao, Yuefan, et al.
Published: (2025)

Theoretical Guarantees for High Order Trajectory Refinement in Generative Flows
by: Gong, Chengyue, et al.
Published: (2025)

HOFAR: High-Order Augmentation of Flow Autoregressive Transformers
by: Liang, Yingyu, et al.
Published: (2025)

High-Order Matching for One-Step Shortcut Diffusion Models
by: Chen, Bo, et al.
Published: (2025)

Universal Approximation of Visual Autoregressive Transformers
by: Chen, Yifang, et al.
Published: (2025)

On Computational Limits and Provably Efficient Criteria of Visual Autoregressive Models: A Fine-Grained Complexity Analysis
by: Ke, Yekun, et al.
Published: (2025)

Visual Autoregressive Transformers Must Use $Ω(n^2 d)$ Memory
by: Cao, Yang, et al.
Published: (2025)

CountDiffusion: Text-to-Image Synthesis with Training-Free Counting-Guidance Diffusion
by: Li, Yanyu, et al.
Published: (2025)

When the Forger Is the Judge: GPT-Image-2 Cannot Recognize Its Own Faked Documents
by: Wu, Jiaqi, et al.
Published: (2026)

Multimodal Language Models Cannot Spot Spatial Inconsistencies
by: Khangaonkar, Om, et al.
Published: (2026)

Prompt Refinement with Image Pivot for Text-to-Image Generation
by: Zhan, Jingtao, et al.
Published: (2024)

Force Matching with Relativistic Constraints: A Physics-Inspired Approach to Stable and Efficient Generative Modeling
by: Cao, Yang, et al.
Published: (2025)

Iterative Prompt Refinement for Safer Text-to-Image Generation
by: Jeon, Jinwoo, et al.
Published: (2025)

Aligned but Stereotypical? The Hidden Influence of System Prompts on Social Bias in LVLM-Based Text-to-Image Models
by: Park, NaHyeon, et al.
Published: (2025)

On Computational Limits of FlowAR Models: Expressivity and Efficiency
by: Cao, Yang, et al.
Published: (2025)

DIR-TIR: Dialog-Iterative Refinement for Text-to-Image Retrieval
by: Zhen, Zongwei, et al.
Published: (2025)

Source Prompt Disentangled Inversion for Boosting Image Editability with Diffusion Models
by: Li, Ruibin, et al.
Published: (2024)

TIP-Editor: An Accurate 3D Editor Following Both Text-Prompts And Image-Prompts
by: Zhuang, Jingyu, et al.
Published: (2024)

Keep What Audio Cannot Say: Context-Preserving Token Pruning for Omni-LLMs
by: Jung, Chaeyoung, et al.
Published: (2026)

Linearized Attention Cannot Enter the Kernel Regime at Any Practical Width
by: Miñoza, Jose Marie Antonio, et al.
Published: (2026)

Contrastive Prompts Improve Disentanglement in Text-to-Image Diffusion Models
by: Wu, Chen, et al.
Published: (2024)

Verify Claimed Text-to-Image Models via Boundary-Aware Prompt Optimization
by: Zhao, Zidong, et al.
Published: (2026)

Anonymization Prompt Learning for Facial Privacy-Preserving Text-to-Image Generation
by: Shi, Liang, et al.
Published: (2024)

Exploring Iterative Refinement with Diffusion Models for Video Grounding
by: Liang, Xiao, et al.
Published: (2023)

Image Super-Resolution with Text Prompt Diffusion
by: Chen, Zheng, et al.
Published: (2023)

HD-Painter: High-Resolution and Prompt-Faithful Text-Guided Image Inpainting with Diffusion Models
by: Manukyan, Hayk, et al.
Published: (2023)

Seek for Incantations: Towards Accurate Text-to-Image Diffusion Synthesis through Prompt Engineering
by: Yu, Chang, et al.
Published: (2024)

Supercharged One-step Text-to-Image Diffusion Models with Negative Prompts
by: Nguyen, Viet, et al.
Published: (2024)

EDITOR: Effective and Interpretable Prompt Inversion for Text-to-Image Diffusion Models
by: Li, Mingzhe, et al.
Published: (2025)

LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models
by: Lian, Long, et al.
Published: (2023)

Debiasing Text-to-Image Diffusion Models
by: He, Ruifei, et al.
Published: (2024)

Dynamic Prompting of Frozen Text-to-Image Diffusion Models for Panoptic Narrative Grounding
by: Li, Hongyu, et al.
Published: (2024)

Local Conditional Controlling for Text-to-Image Diffusion Models
by: Zhao, Yibo, et al.
Published: (2023)

VLM Judges Can Rank but Cannot Score: Task-Dependent Uncertainty in Multimodal Evaluation
by: Kumar, Divake, et al.
Published: (2026)