Neural Network
A computational model loosely inspired by the brain, made of interconnected layers of weighted nodes.
Full Definition
A neural network is a machine learning model composed of layers of interconnected nodes (neurons), each performing a weighted sum of inputs followed by a nonlinear activation function. During training, the weights are adjusted via backpropagation and gradient descent to minimise a loss function on a labelled dataset. Deep neural networks (DNNs) stack many layers, allowing them to learn hierarchical representations — edges → shapes → objects in vision; characters → words → syntax → semantics in language. Modern LLMs are transformer-based deep neural networks with billions of parameters. Understanding neural network fundamentals is prerequisite knowledge for understanding why LLMs behave as they do.
Examples
A three-layer feedforward network learning to classify handwritten digits (MNIST) with 98% accuracy.
A transformer neural network learning to predict the next word in a sentence by adjusting 175 billion weights over trillions of training examples.
Apply this in your prompts
PromptITIN automatically uses techniques like Neural Network to build better prompts for you.
Related Terms
Transformer
The neural network architecture that underpins all modern large language models,…
View →Attention Mechanism
The core transformer operation that weighs the relevance of each token to every …
View →Embedding
A dense numerical vector that represents a token, sentence, or document in a con…
View →