What is a Large Language Model (LLM)?

A Large Language Model, abbreviated LLM, is a neural network trained on large volumes of text to predict the next token in a sequence. Modern LLMs handle a wide range of language tasks including summarization, translation, classification, question answering, and content generation, and they serve as the engine behind most current AI assistants.

Definition

LLMs learn statistical patterns of language by processing billions of tokens during training. Their behavior is shaped further by instruction tuning, reinforcement learning from human feedback, and runtime techniques such as prompting and retrieval. They can be accessed as a service, self-hosted, or fine-tuned on domain-specific data when accuracy or privacy demands it.

Why it matters

LLMs lower the cost of working with unstructured content. In commerce, they draft product descriptions, classify support tickets, answer shopper questions, translate catalogs, and summarize reviews. They are also the conversational layer behind agentic experiences, where the model interprets intent and orchestrates calls to commerce APIs.

Limitations

LLMs are probabilistic and can produce plausible but incorrect output, a behavior called hallucination. Production systems mitigate this through retrieval-augmented generation, structured outputs, guardrails, and human review for high-stakes decisions. They also need careful data handling, because prompts and responses may contain sensitive customer information that has to be protected accordingly.

Related

Explore Agentic Frontend Management Platform · Composable Digital Experience Platform.