A large language model is a neural network, typically based on the transformer architecture, trained on a very large corpus of text to predict the next token in a sequence and able to generate, summarise, classify and reason over natural language.
In practiceLarge language models underpin most current generative AI products. They are characterised by parameter count, training data scale, context window and the alignment techniques applied after pre-training, such as instruction tuning and reinforcement learning from human feedback. Common confusions include treating an LLM as a database, which it is not, and assuming that scale alone solves reliability, when retrieval, tool use, evaluation and human oversight are typically required for production use.
An insurer uses a large language model behind a claims handler assistant: the model drafts a structured claim summary from unstructured notes, while business rules and a human reviewer remain in the loop before any payment decision.
This definition is maintained by Moweb partners and used in live client engagements. For how Large language model applies to your estate, or to challenge a working definition, speak to a partner.