大语言模型（LLM）

A deep learning model pretrained on massive text, good at understanding and generating natural language.

LLMs learn statistical patterns from text corpora and can handle Q&A, writing, translation and code generation. GPT, Claude, Gemini and DeepSeek are examples. Capabilities scale with parameters and data, but they still suffer from hallucination and knowledge cutoff.