Foundation Concepts / Core Definitions

LLM

Beginner [2/5]
Large Language Model Generative language model Neural language model

Definition

A Large Language Model (LLM) is an AI system trained on massive amounts of text data to understand and generate human-like text. These models learn patterns, facts, and reasoning abilities from billions of words of text.

LLMs are the foundation of modern AI assistants like ChatGPT, Claude, and Gemini. They can perform a wide variety of tasks including writing, coding, analysis, translation, and more.

Key Concepts

  • Parameters: The learned weights in the model (billions for large models)
  • Training data: The text corpus the model learned from
  • Inference: The process of generating output from the model
  • Context window: How much text the model can consider at once

Examples

Popular LLMs
Examples of Large Language Models
• GPT-4 (OpenAI) - Powers ChatGPT • Claude (Anthropic) - Known for safety and helpfulness • Gemini (Google) - Multimodal capabilities • Llama (Meta) - Open-source model family • Mistral - Efficient open-source models
Different companies have developed their own LLMs with varying capabilities and strengths.
Model Sizes
Understanding Scale
Small: ~7B parameters (can run on consumer hardware) Medium: ~70B parameters (requires powerful GPUs) Large: ~175B+ parameters (requires data centers) More parameters generally = more capable, but also: • More expensive to run • Slower inference • Higher memory requirements
Model size affects both capability and practical deployment considerations.

Interactive Exercise

🔬
Explore LLM Capabilities

LLMs can handle remarkably diverse tasks. List 5 different types of tasks you could ask an LLM to perform:

Think beyond simple Q&A - consider creative, analytical, and technical tasks.

Pro Tips
  • LLMs are trained on data up to a cutoff date - they don't know recent events
  • They can be confidently wrong - always verify important facts
  • Different LLMs have different strengths and weaknesses
  • The same LLM can give different answers to the same prompt

Related Terms

Prompt Token Context Window Temperature Inference