mlsanityData to 7 Oct 2026

Person-Centric Annotations of LAION-400M: Auditing Bias and Its Transfer to Models

arXiv 2510.03721 ↗First posted 4 Oct 2025cs.CV cs.CL cs.CYcited by 4

On the map: Multimodal LLMs / vision-language models → Social bias and fairness in VLMs

Abstract (arXiv)

Vision-language models trained on large-scale multimodal datasets show strong demographic biases, but the role of training data in producing these biases remains unclear. A major barrier has been the lack of demographic annotations in web-scale datasets such as LAION-400M. We address this gap by creating person-centric annotations for the full dataset, including over 276 million bounding boxes, perceived gender and race/ethnicity labels, and automatically generated captions. These annotations are produced through validated automatic labeling pipelines combining object detection, multimodal captioning, and finetuned classifiers. Using them, we uncover demographic imbalances and harmful associations, such as the disproportionate linking of men and individuals perceived as Black or Middle Eastern with crime-related and negative content. We also show that a linear fit predicts 60-70% of gender bias in CLIP and Stable Diffusion from direct co-occurrences in the data. Our resources establish the first large-scale empirical link between dataset composition and downstream model bias. Code is available at https://github.com/ExplainableML/LAION-400M-Person-Centric-Annotations.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Placed in Social bias and fairness in VLMs, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 103 of its 141 references that are in the map. First posted on arXiv 4 Oct 2025.

Builds on

103 of 141 references.

Closest ancestors · 10

Same direction, newest first

  1. 2024Identifying Implicit Social Biases in Vision-Language Models1 Nov
  2. 2024Social Perception of Faces in a Vision-Language Model26 Aug
  3. 2024SANER: Annotation-free Societal Attribute Neutralizer for Debiasing CLIP19 Aug
  4. 2024Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals30 May
  5. 2023VisoGender: A dataset for benchmarking gender bias in image-text pronoun resolution21 Jun
  6. 2023Balancing the Picture: Debiasing Vision-Language Datasets with Synthetic Contrast Sets24 May
  7. 2023DeAR: Debiasing Vision-Language Models with Additive Residuals18 Mar
  8. 2023MultiModal Bias: Introducing a Framework for Stereotypical Bias Assessment beyond Gender and Race in Vision Language Models16 Mar
  9. 2023Debiasing Vision-Language Models via Biased Prompts31 Jan
  10. 2022A Prompt Array Keeps the Bias Away: Debiasing Vision-Language Models with Adversarial Learning22 Mar

Other references · 93

  1. 2017Attention Is All You Needcited by 59,932
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  5. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  6. 2023Segment Anythingcited by 8,158
  7. 2023Visual Instruction Tuningcited by 7,792
  8. 2020Scaling Laws for Neural Language Modelscited by 6,411
  9. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,322
  10. 2025Qwen2.5-VL Technical Reportcited by 5,418
  11. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  12. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,048
38 more
  1. 2022Training Compute-Optimal Large Language Modelscited by 2,927
  2. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,883
  3. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,769
  4. 2021LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairscited by 1,544
  5. 2015Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Modelscited by 1,352
  6. 2023Kosmos-2: Grounding Multimodal Large Language Models to the Worldcited by 1,075
  7. 2023ShareGPT4V: Improving Large Multi-Modal Models with Better Captionscited by 1,000
  8. 2016Semantics derived automatically from language corpora contain human-like biasescited by 894
  9. 2023Shikra: Unleashing Multimodal LLM's Referential Dialogue Magiccited by 891
  10. 2023EVA-CLIP: Improved Training Techniques for CLIP at Scalecited by 786
  11. 2022PaLI: A Jointly-Scaled Multilingual Language-Image Modelcited by 762
  12. 2023DataComp: In search of the next generation of multimodal datasetscited by 635
  13. 2019Green AIcited by 620
  14. 2024Scaling and evaluating sparse autoencoderscited by 530
  15. 2019Scaling Out-of-Distribution Detection for Real-World Settingscited by 465
  16. 2022Scaling Language-Image Pre-training via Maskingcited by 317
  17. 2023GPT4RoI: Instruction Tuning Large Language Model on Region-of-Interestcited by 300
  18. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  19. 2020Large image datasets: A pyrrhic win for computer vision?cited by 237
  20. 2018Examining Gender and Race Bias in Two Hundred Sentiment Analysis Systemscited by 234
  21. 2023Stable Bias: Analyzing Societal Representations in Diffusion Modelscited by 225
  22. 2023Improving CLIP Training with Language Rewritescited by 216
  23. 2023What does CLIP know about a red circle? Visual prompt engineering for VLMscited by 215
  24. 2019SemEval-2017 Task 4: Sentiment Analysis in Twittercited by 209
  25. 2021General Facial Representation Learning in a Visual-Linguistic Mannercited by 193
  26. 2021Visually Grounded Reasoning across Languages and Culturescited by 188
  27. 2023ViP-LLaVA: Making Large Multimodal Models Understand Arbitrary Visual Promptscited by 160
  28. 2019Towards Fairer Datasets: Filtering and Balancing the Distribution of the People Subtree in the ImageNet Hierarchycited by 156
  29. 2019Does Object Recognition Work for Everyone?cited by 149
  30. 2023Fair Diffusion: Instructing Text-to-Image Generation Models on Fairnesscited by 149
  31. 2021Evaluating CLIP: Towards Characterization of Broader Capabilities and Downstream Implicationscited by 128
  32. 2024BatchTopK Sparse Autoencoderscited by 126
  33. 2019Towards a Critical Race Methodology in Algorithmic Fairnesscited by 124
  34. 2020Image Representations Learned With Unsupervised Pre-Training Contain Human-like Biasescited by 107
  35. 2024Draw-and-Understand: Leveraging Visual Prompts to Enable MLLMs to Comprehend What You Wantcited by 98
  36. 2023Improving Multimodal Datasets with Image Captioningcited by 93
  37. 2024Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discoverycited by 91
  38. 2021Are Gender-Neutral Queries Really Gender-Neutral? Mitigating Gender Bias in Image Searchcited by 90

50 most cited of 93 All 93 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 3.

4 papers.

  1. 2026Tracing Moral Foundations in Large Language Modelscited by 4
  2. 2026Prototypicality Bias Reveals Blindspots in Multimodal Evaluation Metricscited by 2
  3. 2025ProxT2I: Efficient Reward-Guided Text-to-Image Generation via Proximal Diffusioncited by 1
  4. 2026Happy Young Women, Grumpy Old Men? Emotion-Driven Demographic Biases in Synthetic Face Generationcited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2024A Unified Framework and Dataset for Assessing Societal Bias in Vision-Language Models
  2. 2025Data Matters Most: Auditing Social Bias in Contrastive Vision Language Models
  3. 2025DemoBias: An Empirical Study to Trace Demographic Biases in Vision Foundation Models
  4. 2025Bias in the Picture: Benchmarking VLMs with Social-Cue News Images and LLM-as-Judge Assessment
  5. 2024VLBiasBench: A Comprehensive Benchmark for Evaluating Bias in Large Vision-Language Model
  6. 2025Hidden Bias in the Machine: Stereotypes in Text-to-Image Models
  7. 2025BBQ-V: Benchmarking Visual Stereotype Bias in Large Multimodal Models
  8. 2024Examining Gender and Racial Bias in Large Vision-Language Models Using a Novel Dataset of Parallel Images