@inproceedings{ghosh-etal-2025-just,
title = "Just a Scratch: Enhancing {LLM} Capabilities for Self-harm Detection through Intent Differentiation and Emoji Interpretation",
author = "Ghosh, Soumitra and
Singh, Gopendra Vikram and
Shambhavi and
Choudhury, Sabarna and
Ekbal, Asif",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-long.1330/",
doi = "10.18653/v1/2025.acl-long.1330",
pages = "27428--27445",
ISBN = "979-8-89176-251-0",
abstract = "Self-harm detection on social media is critical for early intervention and mental health support, yet remains challenging due to the subtle, context-dependent nature of such expressions. Identifying self-harm intent aids suicide prevention by enabling timely responses, but current large language models (LLMs) struggle to interpret implicit cues in casual language and emojis. This work enhances LLMs' comprehension of self-harm by distinguishing intent through nuanced language{--}emoji interplay. We present the \textit{C}entennial \textit{E}moji \textit{S}ensitivity \textit{M}atrix (\textit{CESM-100}){---}a curated set of 100 emojis with contextual self-harm interpretations{---}and the \textit{S}elf-\textit{H}arm \textit{I}dentification a\textit{N}d intent \textit{E}xtraction with \textit{S}upportive emoji sensitivity (\textit{SHINES}) dataset, offering detailed annotations for self-harm labels, casual mentions (CMs), and serious intents (SIs). Our unified framework:a) enriches inputs using CESM-100;b) fine-tunes LLMs for multi-task learning{---}self-harm detection (primary) and CM/SI span detection (auxiliary);c) generate explainable rationales for self-harm predictions. We evaluate the framework on three state-of-the-art LLMs{---}Llama 3, Mental-Alpaca, and MentalLlama{---}across zero-shot, few-shot, and fine-tuned scenarios. By coupling intent differentiation with contextual cues, our approach commendably enhances LLM performance in both detection and explanation tasks, effectively addressing the inherent ambiguity in self-harm signals. The \textit{SHINES} dataset, \textit{CESM-100} and codebase are publicly available at: https://www.iitp.ac.in/{\%}7eai-nlp-ml/resources.html{\#}SHINES"
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="ghosh-etal-2025-just">
<titleInfo>
<title>Just a Scratch: Enhancing LLM Capabilities for Self-harm Detection through Intent Differentiation and Emoji Interpretation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Soumitra</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gopendra</namePart>
<namePart type="given">Vikram</namePart>
<namePart type="family">Singh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name>
<namePart>Shambhavi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sabarna</namePart>
<namePart type="family">Choudhury</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Asif</namePart>
<namePart type="family">Ekbal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wanxiang</namePart>
<namePart type="family">Che</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joyce</namePart>
<namePart type="family">Nabende</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ekaterina</namePart>
<namePart type="family">Shutova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mohammad</namePart>
<namePart type="given">Taher</namePart>
<namePart type="family">Pilehvar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vienna, Austria</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-251-0</identifier>
</relatedItem>
<abstract>Self-harm detection on social media is critical for early intervention and mental health support, yet remains challenging due to the subtle, context-dependent nature of such expressions. Identifying self-harm intent aids suicide prevention by enabling timely responses, but current large language models (LLMs) struggle to interpret implicit cues in casual language and emojis. This work enhances LLMs’ comprehension of self-harm by distinguishing intent through nuanced language–emoji interplay. We present the Centennial Emoji Sensitivity Matrix (CESM-100)—a curated set of 100 emojis with contextual self-harm interpretations—and the Self-Harm Identification aNd intent Extraction with Supportive emoji sensitivity (SHINES) dataset, offering detailed annotations for self-harm labels, casual mentions (CMs), and serious intents (SIs). Our unified framework:a) enriches inputs using CESM-100;b) fine-tunes LLMs for multi-task learning—self-harm detection (primary) and CM/SI span detection (auxiliary);c) generate explainable rationales for self-harm predictions. We evaluate the framework on three state-of-the-art LLMs—Llama 3, Mental-Alpaca, and MentalLlama—across zero-shot, few-shot, and fine-tuned scenarios. By coupling intent differentiation with contextual cues, our approach commendably enhances LLM performance in both detection and explanation tasks, effectively addressing the inherent ambiguity in self-harm signals. The SHINES dataset, CESM-100 and codebase are publicly available at: https://www.iitp.ac.in/%7eai-nlp-ml/resources.html#SHINES</abstract>
<identifier type="citekey">ghosh-etal-2025-just</identifier>
<identifier type="doi">10.18653/v1/2025.acl-long.1330</identifier>
<location>
<url>https://aclanthology.org/2025.acl-long.1330/</url>
</location>
<part>
<date>2025-07</date>
<extent unit="page">
<start>27428</start>
<end>27445</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Just a Scratch: Enhancing LLM Capabilities for Self-harm Detection through Intent Differentiation and Emoji Interpretation
%A Ghosh, Soumitra
%A Singh, Gopendra Vikram
%A Choudhury, Sabarna
%A Ekbal, Asif
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%A Shambhavi
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F ghosh-etal-2025-just
%X Self-harm detection on social media is critical for early intervention and mental health support, yet remains challenging due to the subtle, context-dependent nature of such expressions. Identifying self-harm intent aids suicide prevention by enabling timely responses, but current large language models (LLMs) struggle to interpret implicit cues in casual language and emojis. This work enhances LLMs’ comprehension of self-harm by distinguishing intent through nuanced language–emoji interplay. We present the Centennial Emoji Sensitivity Matrix (CESM-100)—a curated set of 100 emojis with contextual self-harm interpretations—and the Self-Harm Identification aNd intent Extraction with Supportive emoji sensitivity (SHINES) dataset, offering detailed annotations for self-harm labels, casual mentions (CMs), and serious intents (SIs). Our unified framework:a) enriches inputs using CESM-100;b) fine-tunes LLMs for multi-task learning—self-harm detection (primary) and CM/SI span detection (auxiliary);c) generate explainable rationales for self-harm predictions. We evaluate the framework on three state-of-the-art LLMs—Llama 3, Mental-Alpaca, and MentalLlama—across zero-shot, few-shot, and fine-tuned scenarios. By coupling intent differentiation with contextual cues, our approach commendably enhances LLM performance in both detection and explanation tasks, effectively addressing the inherent ambiguity in self-harm signals. The SHINES dataset, CESM-100 and codebase are publicly available at: https://www.iitp.ac.in/%7eai-nlp-ml/resources.html#SHINES
%R 10.18653/v1/2025.acl-long.1330
%U https://aclanthology.org/2025.acl-long.1330/
%U https://doi.org/10.18653/v1/2025.acl-long.1330
%P 27428-27445
Markdown (Informal)
[Just a Scratch: Enhancing LLM Capabilities for Self-harm Detection through Intent Differentiation and Emoji Interpretation](https://aclanthology.org/2025.acl-long.1330/) (Ghosh et al., ACL 2025)
ACL