LLM

An LLM is a neural network trained on massive text datasets to understand and generate human language. LLMs power chatbots, code assistants, and content generation tools by predicting the most likely next tokens in a sequence.

Scale in parameters and training data are the key drivers of emergent capabilities including reasoning, instruction following, and tool use.