# Muslim-Violence Bias Persists in Debiased GPT Models

arXiv: https://arxiv.org/abs/2310.18368 · HTML: https://mlsanity.com/paper/2310.18368/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 8 references that are in the map.
- First posted on arXiv 25 Oct 2023.

## Builds on (most cited first, 5 of 5)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey](https://mlsanity.com/paper/2111.01243/) (2021)
- [Fairness in representation: quantifying stereotyping as a representational harm](https://mlsanity.com/paper/1901.09565/) (2019)
- [Pretrained AI Models: Performativity, Mobility, and Change](https://mlsanity.com/paper/1909.03290/) (2019)

## Built on by (most cited first, 4 of 4)

- [Biases in the Blind Spot: Detecting What LLMs Fail to Mention](https://mlsanity.com/paper/2602.10117/) (2026)
- [With a Grain of SALT: Are LLMs Fair Across Social Dimensions?](https://mlsanity.com/paper/2410.12499/) (2024)
- [Omissive Bias in Religious Representation: Benchmarking LLM Answers to Everyday Ethical Decision-making](https://mlsanity.com/paper/2605.24319/) (2026)
- [Is Lying Only Sinful in Islam? Exploring Religious Bias in Multilingual Large Language Models Across Major Religions](https://mlsanity.com/paper/2512.03943/) (2025)

## Abstract (arXiv)

Abid et al. (2021) showed a tendency in GPT-3 to generate mostly violent completions when prompted about Muslims, compared with other religions. Two pre-registered replication attempts found few violent completions and only a weak anti-Muslim bias in the more recent InstructGPT, fine-tuned to eliminate biased and toxic outputs. However, more pre-registered experiments showed that using common names associated with the religions in prompts increases several-fold the rate of violent completions, revealing a significant second-order anti-Muslim bias. ChatGPT showed a bias many times stronger regardless of prompt format, suggesting that the effects of debiasing were reduced with continued model development. Our content analysis revealed religion-specific themes containing offensive stereotypes across all experiments. Our results show the need for continual de-biasing of models in ways that address both explicit and higher-order associations.