Filter UNSAFE prompts with Llama Guard 3
Content moderation for Large Language Models (LLMs) involves the detection and filtering of harmful or unwanted content generated by these models. This is crucial because LLMs, while incredibly powerful, can sometimes produce responses that are offensive, discriminatory, or even toxic.
Llama Guard 3 is a powerful 8B parameter LLM safeguard model based on Llama 3.1-8B. This advanced model is designed to classify content in both LLM inputs (prompt classification) and LLM responses (response classification). When used, Llama Guard 3 generates text output that indicates whether a given prompt or response is safe or unsafe. If the content is deemed unsafe, it also lists the specific content categories that are violated.
In this video we will see how we can use Llama Guard 3 with @GroqInc python libraries to flag unsafe content and prevent it from going forward to LLM for response.
Thank you for watching please like share and subscribe!!
0:00 Introduction
0:17 What is Content moderation
1:45 How to use Llama Guard 3
3:42 Test Llama Guard 3 with Groq cloud playground
5:31 Using Llama Guard 3 with Groq python client
7:52 Conclusion
8:07 Like Share and subscribe
Colab notebook - https://colab.research.google.com/gis...
Llama Guard 3 - https://console.groq.com/docs/content...
Join this channel to get access to perks:
/ @superlazycoder1984
Have an early access to videos and Super Coder badge