The AI Glossary

An AI system that can take actions on its own, such as using tools, browsing, or running code, to complete a multi-step task rather than just answering.

The effort to make AI systems behave in line with human intentions and values.

An application programming interface: a way for software to talk to an AI model programmatically instead of through a chat window.

A hypothetical AI that can match or exceed humans across essentially any intellectual task.

A standardized test used to measure and compare AI models on tasks like coding, math, or reasoning.

Prompting or training a model to reason step by step before giving a final answer.

The maximum amount of text (measured in tokens) a model can consider at once, including your prompt and its reply.

The type of AI model behind most image and video generators. It creates images by gradually removing noise until a picture forms.

Training a smaller model to copy the behavior of a larger one, keeping much of the quality at lower cost.

A list of numbers that represents the meaning of a piece of text, so that similar meanings have similar numbers.

Further training an existing model on your own examples so it specializes in a specific task or style.

A large, general-purpose model trained on broad data that can be adapted to many downstream tasks.

Specialized chips (graphics processing units and tensor processing units) that do the heavy math for training and running AI.

The rules and filters that keep an AI from producing harmful, unsafe, or off-limits output.

When an AI states something false or made-up with total confidence.

The act of running a trained model to get an output, as opposed to training it.

A prompt crafted to trick an AI into ignoring its safety rules and producing restricted output.

An AI model trained on huge amounts of text to predict and generate language. LLMs power chatbots like ChatGPT and Claude.

An open standard that lets AI models connect to external tools and data sources in a consistent way.

An AI model that can work with more than one type of input or output, such as text, images, audio, and video.

A model whose trained weights are released publicly so anyone can run, inspect, or fine-tune it.

The internal values a model learns during training. Model size is often described by parameter count.

The instruction or question you give an AI model to get a response.

The practice of writing and refining prompts to get better results from an AI model.

An attack where hidden instructions in content the AI reads hijack its behavior.

Shrinking a model by storing its numbers at lower precision, which reduces memory and speeds it up with a small quality trade-off.

A model designed to spend extra time thinking through a problem before answering, often trading speed for accuracy on hard tasks.

A method where an AI looks up relevant information from your documents and includes it in the prompt before answering.

Reinforcement Learning from Human Feedback: training a model to prefer answers that human reviewers rated as better.

Data generated by AI rather than collected from the real world, used to train or improve other models.

A setting that controls how random or creative a model's output is. Low temperature is focused and predictable; high is more varied.

The small chunk of text an AI model reads and writes. A token is often a word or part of a word, roughly four characters of English on average.

The process of teaching a model by showing it large amounts of data and adjusting its internal settings.

The neural network architecture behind virtually every modern large language model, introduced by Google researchers in 2017.

A database built to store embeddings and quickly find the ones most similar to a query.

Zero-shot means asking a model to do a task with no examples; few-shot means giving it a handful of examples in the prompt first.