Reproduce “Fake News Detection and Classification: A Comparative Study of Convolutional Neural Networks, Large Language Models, and Natural Language Processing Models” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Konstantinos I. Roumeliotis, Nikolaos D. Tselikas, Dimitrios K. Nasiopoulos
The authors test whether fine-tuning GPT-4o-mini improves binary fake-news classification on a held-out WELFake dataset. They train it on 3,200 labeled articles and evaluate it on 1,000 articles capped at 2,560 characters. Its accuracy rises from 24.3% to 98.6%, measuring agreement with dataset labels on this split, not verification of real-world truth or performance beyond it.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.