Getting full potential out of LLM projects

Hosted by Eghed, with Kits.

Are you eager to explore the transformative potential of Large Language Models (LLMs) in your projects? Join us for a hands-on workshop tailored specifically for those new to the world of LLMs. We will go through the fundamentals of LLMs and the tools and know-how to get the full potential out of pre-trained models. The workshop aims to cast a wide net rather than go in-depth on a small sub-problem. We offer a beginner-friendly overview of key concepts and practical applications.

Who is it for?

This workshop is perfect for people curious about integrating LLMs into their workflows but with limited or no prior experience with these models. It is designed to provide a gentle yet insightful introduction to the world of LLMs.

What this workshop offers

In this workshop, participants will:

1. Gain an introduction to LLMs: Explore the foundational concepts of LLMs, including embeddings, model variants, and considerations for selecting the suitable pre-trained model for your task.
2. Learn prompt engineering techniques: Discover how to craft effective prompts to guide LLMs in generating desired outputs, with practical examples showcasing successful and ineffective prompts.
3. Explore the RAG framework: Dive into the innovative Retrieval-Augmented Generation (RAG) framework, uncovering its potential to enhance text completion and semantic search tasks with minimal complexity.
4. Introduction to fine-tuning: Get acquainted with basic LLM fine-tuning for custom tasks, including an overview of popular techniques.

Prerequisites

No prior experience with LLMs is required. However, participants should have:

• A basic understanding of machine learning concepts and algorithms.
• Proficiency in the Python programming language and familiarity with Jupyter notebooks.
• Curiosity and eagerness to learn about NLP concepts and applications.
• A laptop for the code-along steps.