See neuron([1,1,1]) produce 0.3
one move: compute a neuron
Open the code pane immediately, then use the Spec tab, optional design question, tests, and artifact when you want them.
Finish each move before carrying the artifact forward.
A neuron is a weighted sum plus bias followed by a nonlinear activation.
Deliverable: Write the mechanism in one sentence.Implement relu(w·x+b) for a small input vector.
Deliverable: Type the smallest runnable implementation.Check shape, negative pre-activation, and a hand-computed forward value.
Deliverable: Record the expected output and one edge case.Swap ReLU for sigmoid and vary one input at a time.
Deliverable: Change one variable and explain the result.A deterministic neuron forward-pass function.
Deliverable: Save the code, output, and a short failure note.Check shape, negative pre-activation, and a hand-computed forward value.
Swap ReLU for sigmoid and vary one input at a time.
A deterministic neuron forward-pass function.
Mapped from the supplied Build Everything PDF, source page 21. The five moves are Deriva’s implementation contract for this project.