Introduction The AI Agent that were are going to build is going to do the following things: 1) When asked to summarize the content of a website, the AI agent will summarize the content of the webs...

Introduction Setting up the AI Agent echo -e "\033[0;32m----------Updating apt package installer, and installing python package manager __pip__----------\033[0m" apt update && apt install ...

Introduction Scope of threat modeling can be the amount of application features or design that needs to be threat modeled, and the time that is available for the threat modeling activity. Traditio...

Understanding Model Signing and Verification Cosign is a powerful tool that simplifies the signing and verification process for various digital artifacts, including container images. Its capabiliti...

Introduction The lab is not about creating a super helpful model or a chatbot, but understanding how to embed executable code inside a model. Requirements mkdir malicious-models cd malicious-model...

Introduction Model editing refers to the process of modifying a pretrained machine learning model’s internal knowledge or behavior without retraining it from scratch. The goal of model editing is ...

Introduction The lab is not about creating a super helpful model or a chatbot, but understanding how to embed executable code inside a model. Requirements mkdir trojan-neural-network-model cd troj...

Introduction The lab is not about creating a super helpful model or a chatbot, but understanding how to embed executable code inside a model. Requirements mkdir trojan-model cd trojan-model apt up...

Introduction Requirements apt update && apt install -y python3-virtualenv mkdir llm-chatbot-with-guard cd llm-chatbot-with-guard virtualenv venv source venv/bin/activate cat > requirem...

Introduction What is Garak? Garak (Generative AI Red-teaming & Assessment Kit) is a comprehensive toolkit designed for security testing of language models. Think of it as the “nmap” for LLMs -...