A survey of textual cyber abuse detection using cutting-edge language models and large language models (arXiv)

The study offers a thorough examination of the many types of abuse that are common on social media, with an emphasis on how new technologies like Language Models (LMs) and Large Language Models (LLMs) are changing how abusive content is found and produced on these networks. We examine the methods used to spread abuse on social media and examine the psychological and societal effects. The dual function of sophisticated language models is also explored, with an emphasis on how they might improve automated systems for detecting abusive conduct while simultaneously recognizing their potential to produce damaging material.

https://arxiv.org/abs/2501.05443