large language model (LLM)
noun

Definition:
A type of advanced artificial intelligence model trained on massive datasets of human language to understand, generate, and manipulate natural language with high fluency and contextual awareness. Large language models are built using deep learning architectures—most commonly transformer-based—and are capable of performing a wide range of language tasks, including text completion, translation, summarization, question answering, coding, and reasoning.

LLMs learn statistical patterns and semantic relationships between words, phrases, and concepts by processing billions to trillions of tokens (text segments), enabling them to generate coherent and contextually relevant responses. Unlike rule-based systems, LLMs do not possess inherent understanding or knowledge, but simulate understanding through pattern recognition and probabilistic prediction.

Key Characteristics:

• Scale: Trained on massive corpora, often comprising web content, books, scientific articles, code, and other textual data.
• Capabilities: Perform zero-shot, few-shot, and fine-tuned tasks without task-specific programming.
• Contextual Understanding: Maintain conversational memory, adapt tone/style, and infer intent based on input sequences.
• Multimodality (in newer models): Some LLMs, like GPT-4o, can process not just text but also images, audio, and video.

Usage:
“Developers integrated a large language model into their support system to automatically answer customer inquiries in natural language.”

Examples:
GPT-4, Claude, Gemini, LLaMA, Mistral, Perplexity

Compare:
Natural Language Processing (NLP), Generative AI, Neural Network, Foundation Model

First Known Use:
2018, popularized by the release of OpenAI’s GPT-2, though the foundational research dates back earlier.

Note:
LLMs raise important discussions around ethics, bias, hallucination (confident but incorrect outputs), data provenance, and the role of AI in shaping public knowledge.