← Curriculum map
L14 · Phase 4 · 20–35 min, self-paced

Mathematics of Reasoning

“I've now seen several small pieces of maths — how do they fit together, and where would I look for the full picture?”

Everything so far, in one place

Not new material — a map of every equation the course has used, and where it came from. Click any lesson to jump back for a refresher.

L03Similarity and the Dot Producta·b = a₁b₁ + a₂b₂ + … + aₙbₙL04Neural Networksy = f(Wx + b)L05How a Network Learnsw ← w − η·∂L/∂wL06Language Modelssoftmax(x)ᵢ = eˣⁱ ÷ Σⱼ eˣʲL08AttentionAttention(Q,K,V) = softmax(QKᵀ/√dₖ)VL09Why Multiple Attention Heads?MultiHead(Q,K,V) = Concat(head₁,…,headₕ)WₒL13Memorization vs Generalization vs ReasoningQ = mcΔT

Depth ladder

This checkpoint is entirely optional — every Everyday explanation you've read so far already stands on its own, and skipping this lesson costs you nothing at that level.

Knowledge check

Match each equation shown (dot product, softmax, gradient descent, Q=mcΔT, attention, multi-head attention) to the lesson that introduced it.