LLMs

Pretraining is the first and largest stage of training a foundation model, in which the model learns general knowledge and language patterns from vast amounts of unlabelled data, usually by predicting the next token.

A language model is an AI model that estimates how likely a sequence of words is, usually by predicting the next word or token. Language models range from simple word-count statistics to today's large neural networks.

A vector embedding is a list of numbers generated by an embedding model to represent a piece of content, such as a sentence, document or image, so that computers can measure how similar two pieces of content are by comparing their vectors.

Parameter-efficient fine-tuning (PEFT) is a family of methods that adapt large pretrained models by training only a small fraction of parameters, keeping most of the original model frozen, which cuts memory, cost and storage.

Fine-tuning is the process of taking a pretrained AI model and training it further on a smaller, specific dataset so it performs better on a particular task, domain or style.

Supervised fine-tuning (SFT) is the process of further training a pretrained model on labelled examples, usually prompts paired with ideal responses, so it learns specific behaviours, formats or tasks.

A reasoning model is a large language model trained to work through problems step by step, generating internal 'thinking' before its final answer. It uses extra computation at inference time to improve accuracy on maths, coding, science and planning.

An AI model checkpoint is a saved snapshot of a model's learned weights, and often its training state, at a specific point in training. Checkpoints let teams resume training, compare versions, share models and deploy them.

A system prompt is a set of instructions given to an AI model, usually by the developer, that defines its role, rules, tone and boundaries for an entire conversation, separate from the user's individual messages.

Temperature is a setting that controls how random or predictable a language model's output is. Low temperature makes the model choose the most likely words; high temperature spreads probability out, producing more varied and creative text.

LLM inference is the process of running a trained large language model to generate a response: processing the prompt, then producing output one token at a time. Its speed and cost depend on model size, context length, hardware and serving techniques.

LoRA (low-rank adaptation) is a parameter-efficient fine-tuning method that freezes a model's original weights and trains two small matrices whose product represents the weight update, dramatically reducing the parameters and memory needed.

A multimodal LLM is a large language model that can process more than text, such as images, audio or video, usually by converting them into token-like representations the language model can reason over alongside text.