Reproduce “Enhancing Phishing Email Identification with Large Language 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

Catherine Lee

This research from Georgia Institute of Technology investigates the use of Large Language Models to enhance phishing email identification through advanced contextual understanding. Llama-3.1-70b achieved 97.21% accuracy in distinguishing phishing from legitimate emails, demonstrating that LLMs can effectively classify threats and provide interpretable explanations for their decisions.

Key insights

  • Large Language Models demonstrated effectiveness in identifying various suspicious elements within emails, including mismatched sender domains, malicious URLs, and subtle linguistic cues.
  • Larger LLM models generally yielded higher accuracy and reliability scores in phishing detection, with Llama-3.1-70b showing the highest performance.
  • The LLM-based approach can provide interpretable reasoning for classifications, which helps users understand why an email was flagged.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.