From the course: Introduction to LLM Vulnerabilities

Unlock this course with a free trial

Join today to access over 24,000 courses taught by industry experts.

Overview of an embedded-model application

Overview of an embedded-model application

From the course: Introduction to LLM Vulnerabilities

Start my 1-month free trial Buy for my team

Overview of an embedded-model application

“

- [Instructor] Now let's see what does it mean to have an embedded model, an application with an embedded model? That means that the model, the large language model, it would be within the application. So I have an example repository here. We have several different files. It is Python. I'm not going to go into details as to what exactly these Python files are doing, but we're going to take a look and poke around the application and see what does it mean to have this model inside. So we have a main.py, a Python file, and I'm going to scroll here. And the first, and most important thing, is this line right here. Now this is a super, super old model, GPT-2. That means that this model is kind of like the great grandfather of GPT-4, which is kind of like what we're seeing in modern applications like ChatGPT. This is text generation and it means that this will dynamically grab that model and put it inside the application when we run it. That's it, like we are going to accept an input that…

Contents