💡 LLMs (Large Language Models)

LLMs (Large Language Models)#

Large neural networks trained on text to understand and generate language.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

LLMs — large language models — are ML models trained on massive text corpora to understand and generate human language. Built on transformer architectures, they handle a wide range of NLP tasks without task-specific training, via zero-shot and few-shot prompting.

How they work#

The architecture is the transformer with self-attention, which lets the model weigh every word in a sequence for context. Training runs over billions to trillions of tokens (books, articles, code, the web) with a deceptively simple objective: predict the next token. From that single objective emerge text generation, summarisation, translation, reasoning and code generation.

The landscape and uses#

The major families include OpenAI’s GPT, Google’s Gemini, Meta’s LLaMA, Anthropic’s Claude and Mistral’s Mixtral. They power conversational assistants, content creation, code copilots, semantic search and retrieval-augmented generation (RAG), and data-science tasks like query-to-SQL.

The limitations#

The caveats are serious: hallucinations (fluent but wrong answers), bias inherited from training data, cost and energy (very high OpEx to train and serve), stale knowledge that lags real-time events, and limited explainability. A useful mental model is a very powerful autocomplete — predict the most likely continuation — that, at sufficient scale, begins to show reasoning-like behaviour.

Theme: AI & ML Concepts · All terminology

Hint

Mind map — connected ideas

AI (Artificial Intelligence) · Machine Learning (ML) · Neural Networks · Embedding · OpenAI API (ML API) · FLOPs

Hint

More in AI & ML Concepts

AI (Artificial Intelligence) · Classification Models · Computer Vision (CV) · Decision Trees · Linear Models · Logistic Regression · Machine Learning (ML) · Medical AI · Natural Language Processing (NLP) · Neural Networks · Regression Models · Support Vector Machines (SVMs) · Target Variable