LLM Overload - A Sponge Attack Framework for AI Disruption

This project was created by Bianka Nagy, Chandler Camarena, Hunor Csapó, Kleon Dósa and Péter Mihály Vörös.

This project was completed as part of the Introduction to Data Security (IDaSec) course. We do not recommend using the provided code or concepts in real-world product applications, as doing so may violate terms of use.

Table of contents

1. Introduction
2. Documentation about self-hosting
3. Where to start
   1. Inputs
      1. Input types
      2. Generating inputs
   2. Attack types, descriptions
   3. Goals and Results
      1. Bianka's attack plan
      2. Chandler's attack plan
      3. Hunor's attack plan
      4. Kleon's attack plan
      5. Peter's attack plan
4. Summary and comparision of results
5. Mitigation and ethical documentation
6. References