LLM (Large Language Model)
A neural network trained on massive text datasets. Examples: GPT-4, Claude, Gemini.
Full definition
LLM (Large Language Model) is a class of neural networks trained on massive text datasets to generate, understand, and transform language. Examples: GPT-4o (OpenAI), Claude 3.5 (Anthropic), Gemini (Google), Llama (Meta).
LLMs operate on the principle of next-token prediction: the model analyzes the input context and generates the most probable continuation. Modern models contain anywhere from 7 billion to 700+ billion parameters.
For SEO practitioners, LLMs matter for two reasons: 1. AI Overviews in Google and Bing are generated by LLMs — content that is legible and structured for these models gains an advantage. 2. Chatbots (ChatGPT, Perplexity) are used as search engine alternatives — this is a new channel for traffic and brand visibility.