An AI model that is trained on fewer parameters than a large language model (LLM); for example, thousands or hundreds of thousands of examples rather than millions and billions. The advantage of a small language model (SLM) is that it can generate more precise results because it is highly tailored to a particular problem. When doing work (inference), the small language model can run more easily in a local device instead of requiring the cloud. Contrast with
large language model. See
inference and
AI data set.