Show in graph
AI

Data → Machine Learning & AI

LLM

A model trained on large text corpora to generate and reason over natural language and code-like sequences.

Motivation

A large language model solves the problem of predicting and generating language-like sequences from learned patterns at scale.

Where it fits

It belongs to AI applications and machine learning systems.

Mental model

An LLM is a trained model that maps context into likely next tokens, enabling conversation, summarization, coding assistance, and reasoning-like behavior.

Important details

Useful systems around LLMs need prompting, retrieval, tool use, safety, evaluation, and monitoring.

Common mistakes

  • Treating an LLM as a database.
  • Ignoring hallucination, evaluation, and product constraints.