Neuro glossary

A neural network is a program that mimics how the brain works. It learns through machine learning and is designed to handle all kinds of tasks.

A model is a specific version of a neural network. A single neural network can have multiple models designed for different tasks.

Chatbot is a neural network capable of answering users' text-based questions in a chat. Examples of such services include the well-known ChatGPT, GigaChat, Yandex's Alice, and other similar alternatives.

A prompt is a user's request to a neural network—the text that helps it understand what needs to be done.

Text-to-image is a type of interaction with a neural network where a text prompt is entered, and an image is generated as the output.

A dataset is a collection of data used to train a particular neural network model. Depending on the dataset, the neural network may have different speech patterns or, in the case of generation, different styles.

A credit (sometimes also called a token) is a form of virtual currency used by various services to access their features.

Gen is a slang term for something that has been created, usually with the help of neural networks.