abbreviation LLM
Large language model
A large language model (LLM) is an AI system trained on massive datasets to predict and generate text, code, or even images. It works by analyzing patterns in data, allowing it to produce coherent and contextually relevant outputs. For users, LLMs change how they interact with technology, enabling faster research, automated content creation, and even coding assistance—but they also raise concerns about accuracy, bias, and ethical use.
For example, an LLM like Mistral AI’s Mixtral 8x7B can process 100,000 words of context, making it useful for summarizing long documents or drafting emails. However, it may occasionally produce incorrect or misleading information (hallucinations), requiring users to verify critical details.
A journalist uses an LLM to quickly draft a news summary from a 20-page report, saving hours of manual work.
In other languages
See also: Generative AI