# Pinpoint Counterfactuals: Reducing social bias in foundation models via localized counterfactual generation

arXiv: https://arxiv.org/abs/2412.09160 · HTML: https://mlsanity.com/paper/2412.09160/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 66 references that are in the map.
- First posted on arXiv 12 Dec 2024.

## Builds on (most cited first, 6 of 6)

- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [A Task is Worth One Word: Learning with Task Prompts for High-Quality Versatile Image Inpainting](https://mlsanity.com/paper/2312.03594/) (2023)
- [HD-Painter: High-Resolution and Prompt-Faithful Text-Guided Image Inpainting with Diffusion Models](https://mlsanity.com/paper/2312.14091/) (2023)
- [Auditing Gender Presentation Differences in Text-to-Image Models](https://mlsanity.com/paper/2302.03675/) (2023)
- [Unsupervised Model Diagnosis](https://mlsanity.com/paper/2410.06243/) (2024)

## Built on by (most cited first, 1 of 1)

- [Counterfactual Reward Model Training for Bias Mitigation in Multimodal Reinforcement Learning](https://mlsanity.com/paper/2508.19567/) (2025)

## Abstract (arXiv)

Foundation models trained on web-scraped datasets propagate societal biases to downstream tasks. While counterfactual generation enables bias analysis, existing methods introduce artifacts by modifying contextual elements like clothing and background. We present a localized counterfactual generation method that preserves image context by constraining counterfactual modifications to specific attribute-relevant regions through automated masking and guided inpainting. When applied to the Conceptual Captions dataset for creating gender counterfactuals, our method results in higher visual and semantic fidelity than state-of-the-art alternatives, while maintaining the performance of models trained using only real data on non-human-centric tasks. Models fine-tuned with our counterfactuals demonstrate measurable bias reduction across multiple metrics, including a decrease in gender classification disparity and balanced person preference scores, while preserving ImageNet zero-shot performance. The results establish a framework for creating balanced datasets that enable both accurate bias profiling and effective mitigation.