@inproceedings{kim-etal-2025-decoding,
title = "Decoding the Rule Book: Extracting Hidden Moderation Criteria from {R}eddit Communities",
author = "Kim, Youngwoo and
Beniwal, Himanshu and
Johnson, Steven L. and
Hartvigsen, Thomas",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.emnlp-main.1034/",
doi = "10.18653/v1/2025.emnlp-main.1034",
pages = "20487--20498",
ISBN = "979-8-89176-332-6",
abstract = "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."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kim-etal-2025-decoding">
<titleInfo>
<title>Decoding the Rule Book: Extracting Hidden Moderation Criteria from Reddit Communities</title>
</titleInfo>
<name type="personal">
<namePart type="given">Youngwoo</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Himanshu</namePart>
<namePart type="family">Beniwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="given">L</namePart>
<namePart type="family">Johnson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Thomas</namePart>
<namePart type="family">Hartvigsen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-332-6</identifier>
</relatedItem>
<abstract>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.</abstract>
<identifier type="citekey">kim-etal-2025-decoding</identifier>
<identifier type="doi">10.18653/v1/2025.emnlp-main.1034</identifier>
<location>
<url>https://aclanthology.org/2025.emnlp-main.1034/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>20487</start>
<end>20498</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Decoding the Rule Book: Extracting Hidden Moderation Criteria from Reddit Communities
%A Kim, Youngwoo
%A Beniwal, Himanshu
%A Johnson, Steven L.
%A Hartvigsen, Thomas
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-332-6
%F kim-etal-2025-decoding
%X 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.
%R 10.18653/v1/2025.emnlp-main.1034
%U https://aclanthology.org/2025.emnlp-main.1034/
%U https://doi.org/10.18653/v1/2025.emnlp-main.1034
%P 20487-20498
Markdown (Informal)
[Decoding the Rule Book: Extracting Hidden Moderation Criteria from Reddit Communities](https://aclanthology.org/2025.emnlp-main.1034/) (Kim et al., EMNLP 2025)
ACL