mlsanityData to 7 Oct 2026

Word-Level Explanations for Analyzing Bias in Text-to-Image Models

arXiv 2306.05500 ↗First posted 3 Jun 2023cs.CL cs.AI cs.CVcited by 8

Not on the map. Why?

Abstract (arXiv)

Text-to-image models take a sentence (i.e., prompt) and generate images associated with this input prompt. These models have created award wining-art, videos, and even synthetic datasets. However, text-to-image (T2I) models can generate images that underrepresent minorities based on race and sex. This paper investigates which word in the input prompt is responsible for bias in generated images. We introduce a method for computing scores for each word in the prompt; these scores represent its influence on biases in the model's output. Our method follows the principle of explaining by removing, leveraging masked language models to calculate the influence scores. We perform experiments on Stable Diffusion to demonstrate that our method identifies the replication of societal stereotypes in generated images.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 10 of its 17 references that are in the map. First posted on arXiv 3 Jun 2023.

Builds on

10 of 17 references.

References · 10

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  3. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  4. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  5. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  6. 2020Explaining by Removing: A Unified Framework for Model Explanationcited by 218
  7. 2021You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detectioncited by 170
  8. 2022Adapting Pretrained Vision-Language Foundational Models to Medical Imaging Domainscited by 85
  9. 2022No Token Left Behind: Explainability-Aided Image Classification and Generationcited by 25
  10. 2019Gender and Racial Diversity in Commercial Brands' Advertising Images on Social Mediacited by 1

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 5; 2025: 2.

8 papers.

  1. 2024Survey of Bias In Text-to-Image Generation: Definition, Evaluation, and Mitigationcited by 66
  2. 2023Stable Diffusion Exposed: Gender Bias from Prompt to Imagecited by 25
  3. 2024Securing the Future of GenAI: Policy and Technologycited by 4
  4. 2024Multi-Group Proportional Representation in Retrievalcited by 2
  5. 2024GradBias: Unveiling Word Influence on Bias in Text-to-Image Generative Modelscited by 2
  6. 2025How Robust is Model Editing after Fine-Tuning? An Empirical Study on Text-to-Image Diffusion Modelscited by 1
  7. 2024Visual Stereotypes of Autism Spectrum in Janus-Pro-7B, DALL-E, Stable Diffusion, SDXL, FLUX, and Midjourneycited by 0
  8. 2025Testing chatbots on the creation of encoders for audio conditioned image generationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Hidden Bias in the Machine: Stereotypes in Text-to-Image Models
  2. 2025Exposing Hidden Biases in Text-to-Image Models via Automated Prompt Search
  3. 2025Do Existing Testing Tools Really Uncover Gender Bias in Text-to-Image Models?
  4. 2024Generated Bias: Auditing Internal Bias Dynamics of Text-To-Image Generative Models
  5. 2024Identifying Implicit Social Biases in Vision-Language Models
  6. 2023Exploring Social Bias in Downstream Applications of Text-to-Image Foundation Models
  7. 2025Deconstructing Bias: A Multifaceted Framework for Diagnosing Cultural and Compositional Inequities in Text-to-Image Generative Models
  8. 2026BAFIS: Dataset + Framework to assess occupational Bias and Human Preference in modern Text-to-image Models