Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Dalmiere, Antony, Auriol, Guillaume, Nicomette, Vincent, Marchand, Pascal
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2506.22515
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Inhaltsangabe:
  • Traditional phishing detection often overlooks psychological manipulation. This study investigates using Large Language Model (LLM) In-Context Learning (ICL) for fine-grained classification of phishing emails based on a taxonomy of 40 manipulation techniques. Using few-shot examples with GPT-4o-mini on real-world French phishing emails (SignalSpam), we evaluated performance against a human-annotated test set (100 emails). The approach effectively identifies prevalent techniques (e.g., Baiting, Curiosity Appeal, Request For Minor Favor) with a promising accuracy of 0.76. This work demonstrates ICL's potential for nuanced phishing analysis and provides insights into attacker strategies.