# ModelCitizens: Representing Community Voices in Online Safety

arXiv: https://arxiv.org/abs/2507.05455 · HTML: https://mlsanity.com/paper/2507.05455/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 50 references that are in the map.
- First posted on arXiv 7 Jul 2025.

## Builds on (most cited first, 25 of 25)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models](https://mlsanity.com/paper/2403.13372/) (2024)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection](https://mlsanity.com/paper/2203.09509/) (2022)
- [Social Bias Frames: Reasoning about Social and Power Implications of Language](https://mlsanity.com/paper/1911.03891/) (2019)
- [Large Scale Crowdsourcing and Characterization of Twitter Abusive Behavior](https://mlsanity.com/paper/1802.00393/) (2018)
- [Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection](https://mlsanity.com/paper/2111.07997/) (2021)
- [Participatory Research for Low-resourced Machine Translation: A Case Study in African Languages](https://mlsanity.com/paper/2010.02353/) (2020)
- [Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset](https://mlsanity.com/paper/2207.00220/) (2022)
- [Participation is not a Design Fix for Machine Learning](https://mlsanity.com/paper/2007.02423/) (2020)
- [NLPositionality: Characterizing Design Biases of Datasets and Models](https://mlsanity.com/paper/2306.01943/) (2023)
- [Toxicity Detection: Does Context Really Matter?](https://mlsanity.com/paper/2006.00998/) (2020)
- [When the Majority is Wrong: Modeling Annotator Disagreement for Subjective Tasks](https://mlsanity.com/paper/2305.06626/) (2023)
- [The Participatory Turn in AI Design: Theoretical Foundations and the Current State of Practice](https://mlsanity.com/paper/2310.00907/) (2023)
- [Is Your Toxicity My Toxicity? Exploring the Impact of Rater Identity on Toxicity Annotation](https://mlsanity.com/paper/2205.00501/) (2022)
- [When Do Annotator Demographics Matter? Measuring the Influence of Annotator Demographics with the POPQUORN Dataset](https://mlsanity.com/paper/2306.06826/) (2023)
- [Exploring Cross-Cultural Differences in English Hate Speech Annotations: From Dataset Construction to Analysis](https://mlsanity.com/paper/2308.16705/) (2023)
- [The Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels](https://mlsanity.com/paper/2405.05860/) (2024)
- [Queer In AI: A Case Study in Community-Led Participatory AI](https://mlsanity.com/paper/2303.16972/) (2023)
- [COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements](https://mlsanity.com/paper/2306.01985/) (2023)
- [Hate Speech and Counter Speech Detection: Conversational Context Does Matter](https://mlsanity.com/paper/2206.06423/) (2022)
- [Participation in the age of foundation models](https://mlsanity.com/paper/2405.19479/) (2024)
- [Comparing Bad Apples to Good Oranges: Aligning Large Language Models via Joint Preference Optimization](https://mlsanity.com/paper/2404.00530/) (2024)
- [HateBench: Benchmarking Hate Speech Detectors on LLM-Generated Content and Hate Campaigns](https://mlsanity.com/paper/2501.16750/) (2025)
- [HateModerate: Testing Hate Speech Detectors against Content Moderation Policies](https://mlsanity.com/paper/2307.12418/) (2023)

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

- [Multi-Objective Alignment of Language Models for Personalized Psychotherapy](https://mlsanity.com/paper/2602.16053/) (2026)

## Abstract (arXiv)

Automatic toxic language detection is critical for creating safe, inclusive online spaces. However, it is a highly subjective task, with perceptions of toxic language shaped by community norms and lived experience. Existing toxicity detection models are typically trained on annotations that collapse diverse annotator perspectives into a single ground truth, erasing important context-specific notions of toxicity such as reclaimed language. To address this, we introduce MODELCITIZENS, a dataset of 6.8K social media posts and 40K toxicity annotations across diverse identity groups. To capture the role of conversational context on toxicity, typical of social media posts, we augment MODELCITIZENS posts with LLM-generated conversational scenarios. State-of-the-art toxicity detection tools (e.g. OpenAI Moderation API, GPT-o4-mini) underperform on MODELCITIZENS, with further degradation on context-augmented posts. Finally, we release LLAMACITIZEN-8B and GEMMACITIZEN-12B, LLaMA- and Gemma-based models finetuned on MODELCITIZENS, which outperform GPT-o4-mini by 5.5% on in-distribution evaluations. Our findings highlight the importance of community-informed annotation and modeling for inclusive content moderation. The data, models and code are available at https://github.com/asuvarna31/modelcitizens.