Neural Networks, Actually Explained

You have heard the term a thousand times. Now let's actually look inside: neurons, weights, activations, and what happens in a single pass.

35 min · Reviewed 2026

The Artificial Neuron

A neural network is built out of tiny units called neurons. Each neuron is simple. It takes in a few numbers, multiplies each by a weight, adds them up, and runs the total through a simple squishing function. That is it.

One neuron, step by step

Receive inputs: x1, x2, x3
Multiply by weights: w1·x1 + w2·x2 + w3·x3
Add a bias number: total + b
Pass through an activation function like ReLU or sigmoid
Send the result to the next neuron

Stacking into layers

A layer is many neurons working at the same time, each with its own weights. Stack many layers in a row and you have a deep neural network. The output of one layer becomes the input to the next.

# A tiny feed-forward network in pseudo-code
h1 = relu(W1 @ x + b1)
h2 = relu(W2 @ h1 + b2)
output = softmax(W3 @ h2 + b3)Three layers: input → hidden → hidden → output. @ means matrix multiplication.

What gets learned

The weights W and biases b in every layer
Nothing else — the structure stays fixed during training
A modern LLM has hundreds of billions of these numbers

Why deep beats shallow

Early layers pick up simple features like edges in images or letter patterns in text. Middle layers combine those into shapes or word parts. Late layers pull it all together into concepts like dog or sarcasm. Depth lets the model build up abstractions.

The network's layers are a ladder from pixels to meaning.
— Yoshua Bengio, paraphrased

The big idea: a neural network is a huge stack of very simple math units, wired together in layers. Depth and scale turn that simplicity into something that can translate languages and write poetry.

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-builders-neural-networks-explained

What is the core idea behind "Neural Networks, Actually Explained"?
1. You have heard the term a thousand times. Now let's actually look inside: neurons, weights, activations, and what happens in a single pass.
2. If something happened after AI's training, AI may not know about it.
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
Which term best describes a foundational idea in "Neural Networks, Actually Explained"?
1. weight
2. neuron
3. bias
4. activation function
A learner studying Neural Networks, Actually Explained would need to understand which concept?
1. neuron
2. bias
3. weight
4. activation function
Which of these is directly relevant to Neural Networks, Actually Explained?
1. neuron
2. weight
3. activation function
4. bias
Which of the following is a key point about Neural Networks, Actually Explained?
1. Receive inputs: x1, x2, x3
2. Multiply by weights: w1·x1 + w2·x2 + w3·x3
3. Add a bias number: total + b
4. Pass through an activation function like ReLU or sigmoid
Which of these does NOT belong in a discussion of Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Receive inputs: x1, x2, x3
3. Add a bias number: total + b
4. Multiply by weights: w1·x1 + w2·x2 + w3·x3
Which statement is accurate regarding Neural Networks, Actually Explained?
1. Nothing else — the structure stays fixed during training
2. A modern LLM has hundreds of billions of these numbers
3. The weights W and biases b in every layer
4. If something happened after AI's training, AI may not know about it.
What is the key insight about "Activation functions" in the context of Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. Guarantee that principles compose without conflict on edge cases
4. Without a squishing function, the network could only compute straight-line relationships.
What is the recommended tip about "Build your mental model" in the context of Neural Networks, Actually Explained?
1. AI isn't magic — it's pattern recognition at scale. The more you understand how it works, the more effectively you can u…
2. If something happened after AI's training, AI may not know about it.
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
Which statement accurately describes an aspect of Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. A neural network is built out of tiny units called neurons. Each neuron is simple.
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
What does working with Neural Networks, Actually Explained typically involve?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. A layer is many neurons working at the same time, each with its own weights. Stack many layers in a row and you have a deep neural network.
4. Guarantee that principles compose without conflict on edge cases
Which of the following is true about Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. Guarantee that principles compose without conflict on edge cases
4. Early layers pick up simple features like edges in images or letter patterns in text. Middle layers combine those into shapes or word parts.
Which best describes the scope of "Neural Networks, Actually Explained"?
1. It focuses on You have heard the term a thousand times. Now let's actually look inside: neurons, weights, activati
2. It is unrelated to foundations workflows
3. It applies only to the opposite beginner tier
4. It was deprecated in 2024 and no longer relevant
Which section heading best belongs in a lesson about Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. One neuron, step by step
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
Which section heading best belongs in a lesson about Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. Stacking into layers
4. Guarantee that principles compose without conflict on edge cases

← Back to interactive lesson

Tendril · Builders · AI Foundations

Neural Networks, Actually Explained

You have heard the term a thousand times. Now let's actually look inside: neurons, weights, activations, and what happens in a single pass.

35 min · Reviewed 2026

The Artificial Neuron

One neuron, step by step

Receive inputs: x1, x2, x3
Multiply by weights: w1·x1 + w2·x2 + w3·x3
Add a bias number: total + b
Pass through an activation function like ReLU or sigmoid
Send the result to the next neuron

Stacking into layers

A layer is many neurons working at the same time, each with its own weights. Stack many layers in a row and you have a deep neural network. The output of one layer becomes the input to the next.

# A tiny feed-forward network in pseudo-code
h1 = relu(W1 @ x + b1)
h2 = relu(W2 @ h1 + b2)
output = softmax(W3 @ h2 + b3)Three layers: input → hidden → hidden → output. @ means matrix multiplication.

What gets learned

The weights W and biases b in every layer
Nothing else — the structure stays fixed during training
A modern LLM has hundreds of billions of these numbers

Why deep beats shallow

The network's layers are a ladder from pixels to meaning.
— Yoshua Bengio, paraphrased

The big idea: a neural network is a huge stack of very simple math units, wired together in layers. Depth and scale turn that simplicity into something that can translate languages and write poetry.

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-builders-neural-networks-explained

What is the core idea behind "Neural Networks, Actually Explained"?
1. You have heard the term a thousand times. Now let's actually look inside: neurons, weights, activations, and what happens in a single pass.
2. If something happened after AI's training, AI may not know about it.
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
Which term best describes a foundational idea in "Neural Networks, Actually Explained"?
1. weight
2. neuron
3. bias
4. activation function
A learner studying Neural Networks, Actually Explained would need to understand which concept?
1. neuron
2. bias
3. weight
4. activation function
Which of these is directly relevant to Neural Networks, Actually Explained?
1. neuron
2. weight
3. activation function
4. bias
Which of the following is a key point about Neural Networks, Actually Explained?
1. Receive inputs: x1, x2, x3
2. Multiply by weights: w1·x1 + w2·x2 + w3·x3
3. Add a bias number: total + b
4. Pass through an activation function like ReLU or sigmoid
Which of these does NOT belong in a discussion of Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Receive inputs: x1, x2, x3
3. Add a bias number: total + b
4. Multiply by weights: w1·x1 + w2·x2 + w3·x3
Which statement is accurate regarding Neural Networks, Actually Explained?
1. Nothing else — the structure stays fixed during training
2. A modern LLM has hundreds of billions of these numbers
3. The weights W and biases b in every layer
4. If something happened after AI's training, AI may not know about it.
What is the key insight about "Activation functions" in the context of Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. Guarantee that principles compose without conflict on edge cases
4. Without a squishing function, the network could only compute straight-line relationships.
What is the recommended tip about "Build your mental model" in the context of Neural Networks, Actually Explained?
1. AI isn't magic — it's pattern recognition at scale. The more you understand how it works, the more effectively you can u…
2. If something happened after AI's training, AI may not know about it.
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
Which statement accurately describes an aspect of Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. A neural network is built out of tiny units called neurons. Each neuron is simple.
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
What does working with Neural Networks, Actually Explained typically involve?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. A layer is many neurons working at the same time, each with its own weights. Stack many layers in a row and you have a deep neural network.
4. Guarantee that principles compose without conflict on edge cases
Which of the following is true about Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. Guarantee that principles compose without conflict on edge cases
4. Early layers pick up simple features like edges in images or letter patterns in text. Middle layers combine those into shapes or word parts.
Which best describes the scope of "Neural Networks, Actually Explained"?
1. It focuses on You have heard the term a thousand times. Now let's actually look inside: neurons, weights, activati
2. It is unrelated to foundations workflows
3. It applies only to the opposite beginner tier
4. It was deprecated in 2024 and no longer relevant
Which section heading best belongs in a lesson about Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. One neuron, step by step
3. Pairing low temperature with structured-output tasks like extraction
4. Guarantee that principles compose without conflict on edge cases
Which section heading best belongs in a lesson about Neural Networks, Actually Explained?
1. If something happened after AI's training, AI may not know about it.
2. Pairing low temperature with structured-output tasks like extraction
3. Stacking into layers
4. Guarantee that principles compose without conflict on edge cases

← Back to interactive lesson