What is LLMs (Large Language Models)

Definition

Large language models (LLMs) are neural networks with billions of parameters trained on massive text corpora to understand and generate human language, enabling a wide range of tasks such as writing, summarising, answering questions, and reasoning through prompting alone.

« Back to Glossary Index

Perform diverse language tasks without task-specific training
Generate fluent, context-aware text from natural-language prompts
Adapt to new tasks through prompting and few-shot examples
Serve as flexible foundations for assistants and copilots

Glossary: Stream Processing

Real World Example

A company deploys an LLM-powered assistant that reads internal documents and answers employee questions in natural language, handling everything from policy lookups to drafting emails through prompting.

FAQs

What can LLMs do?

Write, summarise, translate, answer questions, generate code, and reason over text, all driven by prompts.

How are LLMs trained?

They are trained on vast text corpora to predict the next token, then often refined with techniques like RLHF.

What are the limitations of LLMs?

They can hallucinate, lack current or private knowledge unless provided, and reflect biases from their training data.