Bias and Cyberbullying Detection and Data Generation with Transformer AI Models and top LLMs

• DOI: 10.20944/preprints202407.0411.v1
• Publication Date: 2024-07-09T06:39:07Z
• Authors: Yulia Kumar; Kuan Huang; Angelo Perez; Guohao Yang; J. Jenny Li; Patricia Morreale; Dov Kruger; Raymond Jiang

Despite significant advancements in Artificial Intelligence (AI) and Large Language Models (LLMs), detecting mitigating bias remains a critical challenge, particularly within social media platforms like X (formerly Twitter) addressing cyberbullying present on them. This research investigates the effectiveness of leading LLMs generating synthetic biased data evaluates proficiency Transformer AI models both authentic contexts. The study involves semantic analysis feature engineering dataset over 48,000 sentences related to collected from Twitter (before it became X). Leveraging state-of-the-art such as ChatGPT-4o, Pi AI, Claude 3 Opus, Gemini-1.5, biased, cyberbullying, neutral were generated deepen understanding human-generated data. including DeBERTa, Longformer, BigBird, HateBERT, MobileBERT, DistilBERT, BERT, RoBERTa, ELECTRA, XLNet initially trained classify subsequently fine-tuned, optimized, quantized for multilabel classification (detecting biases cyberbullying). study's outcomes include prototype hybrid application that combines Bias Data Detector Generator, validated through extensive testing.