@inproceedings{ghimire-etal-2026-linus,
title = "{LINUS}@{EEUCA} 2026: Fine-grained Toxicity Detection in Gaming Chat using Multilingual Transformers",
author = "Ghimire, Prajwal and
Mahato, Aashish and
Regmi, Sunil",
editor = {H{\"u}rriyeto{\u{g}}lu, Ali and
Thapa, Surendrabikram and
Tanev, Hristo},
booktitle = "Proceedings of the 9th Workshop on Event Extraction and Understanding: Challenges and Applications ({EEUCA} 2026)",
month = jul,
year = "2026",
address = "San Diego, California, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.eeuca-1.24/",
pages = "216--222",
ISBN = "979-8-89176-402-6",
abstract = "The detection of toxic behavior in online gaming communities is crucial for maintaining safe digital spaces, yet remains challenging due to subtle context-dependent and intent-driven language. The GameTox dataset consists of around 53K World of Tanks chat utterances annotated across six categories: Non-toxic, Insults and Flaming, Other Offensive Texts, Hate and Harassment, Threats, and Extremism (CITATION). Our best performing approach, across multiple transformer-based architecture experimentations, is based on the multilingual BERT variant mmBERT-base fine-tuned with class-weighted cross-entropy loss. The best mmBERT-base model achieved a Macro F1 of 0.5882 during validation and an official test Macro F1 of 0.5104 on the shared task leaderboard. An internal held-out evaluation on a development split yielded 0.4282, which we analyze to understand distributional sensitivity to gaming slang and class imbalance. The code is available at: \url{https://github.com/sunilRegmi-ai/eeuca-toxicity-detection}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ghimire-etal-2026-linus">
<titleInfo>
<title>LINUS@EEUCA 2026: Fine-grained Toxicity Detection in Gaming Chat using Multilingual Transformers</title>
</titleInfo>
<name type="personal">
<namePart type="given">Prajwal</namePart>
<namePart type="family">Ghimire</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aashish</namePart>
<namePart type="family">Mahato</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunil</namePart>
<namePart type="family">Regmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 9th Workshop on Event Extraction and Understanding: Challenges and Applications (EEUCA 2026)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ali</namePart>
<namePart type="family">Hürriyetoğlu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Surendrabikram</namePart>
<namePart type="family">Thapa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hristo</namePart>
<namePart type="family">Tanev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, USA</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-402-6</identifier>
</relatedItem>
<abstract>The detection of toxic behavior in online gaming communities is crucial for maintaining safe digital spaces, yet remains challenging due to subtle context-dependent and intent-driven language. The GameTox dataset consists of around 53K World of Tanks chat utterances annotated across six categories: Non-toxic, Insults and Flaming, Other Offensive Texts, Hate and Harassment, Threats, and Extremism (CITATION). Our best performing approach, across multiple transformer-based architecture experimentations, is based on the multilingual BERT variant mmBERT-base fine-tuned with class-weighted cross-entropy loss. The best mmBERT-base model achieved a Macro F1 of 0.5882 during validation and an official test Macro F1 of 0.5104 on the shared task leaderboard. An internal held-out evaluation on a development split yielded 0.4282, which we analyze to understand distributional sensitivity to gaming slang and class imbalance. The code is available at: https://github.com/sunilRegmi-ai/eeuca-toxicity-detection.</abstract>
<identifier type="citekey">ghimire-etal-2026-linus</identifier>
<location>
<url>https://aclanthology.org/2026.eeuca-1.24/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>216</start>
<end>222</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T LINUS@EEUCA 2026: Fine-grained Toxicity Detection in Gaming Chat using Multilingual Transformers
%A Ghimire, Prajwal
%A Mahato, Aashish
%A Regmi, Sunil
%Y Hürriyetoğlu, Ali
%Y Thapa, Surendrabikram
%Y Tanev, Hristo
%S Proceedings of the 9th Workshop on Event Extraction and Understanding: Challenges and Applications (EEUCA 2026)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, USA
%@ 979-8-89176-402-6
%F ghimire-etal-2026-linus
%X The detection of toxic behavior in online gaming communities is crucial for maintaining safe digital spaces, yet remains challenging due to subtle context-dependent and intent-driven language. The GameTox dataset consists of around 53K World of Tanks chat utterances annotated across six categories: Non-toxic, Insults and Flaming, Other Offensive Texts, Hate and Harassment, Threats, and Extremism (CITATION). Our best performing approach, across multiple transformer-based architecture experimentations, is based on the multilingual BERT variant mmBERT-base fine-tuned with class-weighted cross-entropy loss. The best mmBERT-base model achieved a Macro F1 of 0.5882 during validation and an official test Macro F1 of 0.5104 on the shared task leaderboard. An internal held-out evaluation on a development split yielded 0.4282, which we analyze to understand distributional sensitivity to gaming slang and class imbalance. The code is available at: https://github.com/sunilRegmi-ai/eeuca-toxicity-detection.
%U https://aclanthology.org/2026.eeuca-1.24/
%P 216-222
Markdown (Informal)
[LINUS@EEUCA 2026: Fine-grained Toxicity Detection in Gaming Chat using Multilingual Transformers](https://aclanthology.org/2026.eeuca-1.24/) (Ghimire et al., EEUCA 2026)
ACL