# StereoTales: A Multilingual Framework for Open-Ended Stereotype Discovery in LLMs

arXiv: https://arxiv.org/abs/2605.10442 · HTML: https://mlsanity.com/paper/2605.10442/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 34 of its 121 references that are in the map.
- First posted on arXiv 11 May 2026.

## Builds on (most cited first, 30 of 34)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [A Survey of Large Language Models](https://mlsanity.com/paper/2303.18223/) (2023)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [OpenAI GPT-5 System Card](https://mlsanity.com/paper/2601.03267/) (2025)
- [Gemini: A Family of Highly Capable Multimodal Models](https://mlsanity.com/paper/2312.11805/) (2023)
- [DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models](https://mlsanity.com/paper/2512.02556/) (2025)
- [Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods](https://mlsanity.com/paper/1804.06876/) (2018)
- [GLM-5: from Vibe Coding to Agentic Engineering](https://mlsanity.com/paper/2602.15763/) (2026)
- [Kimi K2.5: Visual Agentic Intelligence](https://mlsanity.com/paper/2602.02276/) (2026)
- [Gender Bias in Coreference Resolution](https://mlsanity.com/paper/1804.09301/) (2018)
- [Towards Measuring the Representation of Subjective Global Opinions in Language Models](https://mlsanity.com/paper/2306.16388/) (2023)
- [Measuring short-form factuality in large language models](https://mlsanity.com/paper/2411.04368/) (2024)
- [Measuring Bias in Contextualized Word Representations](https://mlsanity.com/paper/1906.07337/) (2019)
- [Measuring and Reducing Gendered Correlations in Pre-trained Models](https://mlsanity.com/paper/2010.06032/) (2020)
- [Long-form factuality in large language models](https://mlsanity.com/paper/2403.18802/) (2024)
- [A Survey on Fairness in Large Language Models](https://mlsanity.com/paper/2308.10149/) (2023)
- [HalluLens: LLM Hallucination Benchmark](https://mlsanity.com/paper/2504.17550/) (2025)
- [Measuring Implicit Bias in Explicitly Unbiased Large Language Models](https://mlsanity.com/paper/2402.04105/) (2024)
- [MBBQ: A Dataset for Cross-Lingual Comparison of Stereotypes in Generative LLMs](https://mlsanity.com/paper/2406.07243/) (2024)
- [AILuminate: Introducing v1.0 of the AI Risk and Reliability Benchmark from MLCommons](https://mlsanity.com/paper/2503.05731/) (2025)
- [Sociodemographic Bias in Language Models: A Survey and Forward Path](https://mlsanity.com/paper/2306.08158/) (2023)
- [First-Person Fairness in Chatbots](https://mlsanity.com/paper/2410.19803/) (2024)
- [CASE-Bench: Context-Aware SafEty Benchmark for Large Language Models](https://mlsanity.com/paper/2501.14940/) (2025)
- [Towards Auditing Large Language Models: Improving Text-based Stereotype Detection](https://mlsanity.com/paper/2311.14126/) (2023)
- [Fairness Testing of Large Language Models in Role-Playing](https://mlsanity.com/paper/2411.00585/) (2024)
- [Evaluating and Mitigating Social Bias for Large Language Models in Open-ended Settings](https://mlsanity.com/paper/2412.06134/) (2024)
- [Bring Your Own Prompts: Use-Case-Specific Bias and Fairness Evaluation for LLMs](https://mlsanity.com/paper/2407.10853/) (2024)
- [What's Not Said Still Hurts: A Description-Based Evaluation Framework for Measuring Social Bias in LLMs](https://mlsanity.com/paper/2502.19749/) (2025)
- [RealHarm: A Collection of Real-World Language Model Application Failures](https://mlsanity.com/paper/2504.10277/) (2025)
- [Mechanistic Interpretability with SAEs: Probing Religion, Violence, and Geography in Large Language Models](https://mlsanity.com/paper/2509.17665/) (2025)

## Abstract (arXiv)

Multilingual studies of social bias in open-ended LLM generation remain limited: most existing benchmarks are English-centric, template-based, or restricted to recognizing pre-specified stereotypes. We introduce StereoTales, a multilingual dataset and evaluation pipeline for systematically studying the emergence of social bias in open-ended LLM generation. The dataset covers 10 languages and 79 socio-demographic attributes, and comprises over 650k stories generated by 23 recent LLMs, each annotated with the socio-demographic profile of the protagonist across 19 dimensions. From these, we apply statistical tests to identify more than 1,500 over-represented associations, which we then rate for harmfulness through both a panel of humans (N = 247) and the same LLMs. We report three main findings. (i) Every model we evaluate emits consequential harmful stereotypes in open-ended generation, regardless of size or capabilities, and these associations are largely shared across providers rather than isolated misbehaviors. (ii) Prompt language strongly shapes which stereotypes appear: rather than transferring as a shared set of biases, harmful associations adapt culturally to the prompt language and amplify bias against locally salient protected groups. (iii) Human and LLM harmfulness judgments are broadly aligned (Spearman ρ=0.62), with disagreements concentrating on specific attribute classes rather than specific providers. To support further analyses, we release the evaluation code and the dataset, including model generations, attribute annotations, and harmfulness ratings.