The study investigates the capability of GPT-3.5 to produce human-like comments on Dutch news articles. It was found that fine-tuned BERT models can easily distinguish between human-written comments and those generated by GPT-3.5. Despite the fluency of the text generated by Large Language Models (LLMs), the study concludes that their ability to create human-like opinionated comments is still limited.
Publication date: 21 Dec 2023
Project Page: https://arxiv.org/abs/2312.13961v1
Paper: https://arxiv.org/pdf/2312.13961
