From Detection to Explanation: Effective Learning Strategies for LLMs in Online Abusive Language Research

TY - CHAP

T1 - From Detection to Explanation: Effective Learning Strategies for LLMs in Online Abusive Language Research

AU - Di Bonaventura, Chiara

AU - Siciliani, Lucia

AU - Basile, Pierpaolo

AU - Merono Penuela, Albert

AU - McGillivray, Barbara

PY - 2025/1

Y1 - 2025/1

N2 - Abusive language detection relies on understanding different levels of intensity, expressiveness and targeted groups, which requires commonsense reasoning, world knowledge and linguistic nuances that evolve over time. Here, we frame the problem as a knowledge-guided learning task, and demonstrate that LLMs' implicit knowledge without an accurate strategy is not suitable for multi-class detection nor explanation generation. We publicly release GLlama Alarm, the knowledge-Guided version of Llama-2 instruction fine-tuned for multi-class abusive language detection and explanation generation. By being fine-tuned on structured explanations and external reliable knowledge sources, our model mitigates bias and generates explanations that are relevant to the text and coherent with human reasoning, with an average 48.76% better alignment with human judgment according to our expert survey.

AB - Abusive language detection relies on understanding different levels of intensity, expressiveness and targeted groups, which requires commonsense reasoning, world knowledge and linguistic nuances that evolve over time. Here, we frame the problem as a knowledge-guided learning task, and demonstrate that LLMs' implicit knowledge without an accurate strategy is not suitable for multi-class detection nor explanation generation. We publicly release GLlama Alarm, the knowledge-Guided version of Llama-2 instruction fine-tuned for multi-class abusive language detection and explanation generation. By being fine-tuned on structured explanations and external reliable knowledge sources, our model mitigates bias and generates explanations that are relevant to the text and coherent with human reasoning, with an average 48.76% better alignment with human judgment according to our expert survey.

KW - hate speech

KW - Large Language Models

KW - text classification

KW - text generation

M3 - Conference paper

BT - Proceedings of the 2025 International Conference on Computational Linguistics (COLING 2025)

CY - Abu Dhabi

ER -