# Decoding the Rule Book: Extracting Hidden Moderation Criteria from Reddit Communities

arXiv: https://arxiv.org/abs/2509.02926 · HTML: https://mlsanity.com/paper/2509.02926/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 23 references that are in the map.
- First posted on arXiv 3 Sep 2025.

## Builds on (most cited first, 13 of 13)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [A Survey on the Explainability of Supervised Machine Learning](https://mlsanity.com/paper/2011.07876/) (2020)
- [HateCheck: Functional Tests for Hate Speech Detection Models](https://mlsanity.com/paper/2012.15606/) (2020)
- [Detecting Community Sensitive Norm Violations in Online Conversations](https://mlsanity.com/paper/2110.04419/) (2021)
- [What Makes it Ok to Set a Fire? Iterative Self-distillation of Contexts and Rationales for Disambiguating Defeasible Social and Moral Situations](https://mlsanity.com/paper/2310.15431/) (2023)
- [ValueScope: Unveiling Implicit Norms and Values via Return Potential Model of Social Interactions](https://mlsanity.com/paper/2407.02472/) (2024)
- [Making Online Communities 'Better': A Taxonomy of Community Values on Reddit](https://mlsanity.com/paper/2109.05152/) (2021)
- [Interpretable Active Learning](https://mlsanity.com/paper/1708.00049/) (2017)
- [Discovering Biases in Information Retrieval Models Using Relevance Thesaurus as Global Explanation](https://mlsanity.com/paper/2410.03584/) (2024)
- [Uncovering the Internet's Hidden Values: An Empirical Study of Desirable Behavior Using Highly-Upvoted Content on Reddit](https://mlsanity.com/paper/2410.13036/) (2024)

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

- [A Survey of Toxicity Detection and Mitigation Strategies for Multilingual Language Models](https://mlsanity.com/paper/2606.25380/) (2026)

## Abstract (arXiv)

Effective content moderation systems require explicit classification criteria, yet online communities like subreddits often operate with diverse, implicit standards. This work introduces a novel approach to identify and extract these implicit criteria from historical moderation data using an interpretable architecture. We represent moderation criteria as score tables of lexical expressions associated with content removal, enabling systematic comparison across different communities. Our experiments demonstrate that these extracted lexical patterns effectively replicate the performance of neural moderation models while providing transparent insights into decision-making processes. The resulting criteria matrix reveals significant variations in how seemingly shared norms are actually enforced, uncovering previously undocumented moderation patterns including community-specific tolerances for language, features for topical restrictions, and underlying subcategories of the toxic speech classification.