Cyberbullying Detection via Aggression-Enhanced Prompting

TL;DR

This paper explores integrating aggression detection as an auxiliary task within large language models to improve cyberbullying detection on social media, demonstrating that enriched prompts with aggression context enhance performance.

Contribution

The study introduces an enriched prompt pipeline that embeds aggression predictions into cyberbullying detection prompts, significantly improving detection accuracy over standard fine-tuning methods.

Findings

Enriched prompt pipeline outperforms LoRA fine-tuning.

Aggression-informed context boosts detection performance.

Multi-task learning results are inconsistent, but prompt augmentation is effective.

Abstract

Detecting cyberbullying on social media remains a critical challenge due to its subtle and varied expressions. This study investigates whether integrating aggression detection as an auxiliary task within a unified training framework can enhance the generalisation and performance of large language models (LLMs) in cyberbullying detection. Experiments are conducted on five aggression datasets and one cyberbullying dataset using instruction-tuned LLMs. We evaluated multiple strategies: zero-shot, few-shot, independent LoRA fine-tuning, and multi-task learning (MTL). Given the inconsistent results of MTL, we propose an enriched prompt pipeline approach in which aggression predictions are embedded into cyberbullying detection prompts to provide contextual augmentation. Preliminary results show that the enriched prompt pipeline consistently outperforms standard LoRA fine-tuning, indicating that aggression-informed context significantly boosts cyberbullying detection. This study highlights the potential of auxiliary tasks, such as aggression detection, to improve the generalisation of LLMs for safety-critical applications on social networks.