Skip to content

Latest commit

 

History

History
15 lines (15 loc) · 460 Bytes

README.md

File metadata and controls

15 lines (15 loc) · 460 Bytes

Llama Attack

Introduction

This repo contains apps and implementations of adversarial attacks against Large Language Models and Vision Language Models.

Implementations

  • Greedy Coordinate Gradient (by nanoGCG)
  • Don't Say No (modified verstion of nanoGCG)
  • Visual Embedding Attack (different images with similar embeddings)

Apps

Apps are created using Streamlit for implementations.

  • GCG app
  • DSN app
  • VisEmb app

Models

  • Llama 3.1
  • Qwen2-VL