How LLMs Work102 · Module II · Lesson 08 of 31
Article · 12 min

Feed-forward networks

The MLP inside the block, and what it does.

Summary

Every transformer layer contains a feed-forward network (FFN) that processes each position independently after attention. It is roughly two-thirds of the parameters and much of the compute. The FFN is where much of the model's stored knowledge lives.

Objectives
  • 01State the FFN's role and shape.
  • 02Explain why FFN width is usually 4x the model dimension.
  • 03Describe the mixture-of-experts variant.
The Lesson

The layer

Two linear layers with a nonlinearity between: FFN(x) = W2 · σ(W1 · x + b1) + b2. σ is typically GeLU, SwiGLU, or ReLU. Hidden dimension is usually 4x the residual stream dimension.

Knowledge storage

Mechanistic-interpretability work (e.g., transformer-circuits.pub) locates much of what the model 'knows' — facts, associations, template completions — in FFN weights. Attention moves information; FFNs store it.

MoE variant

Mixture-of-experts replaces a dense FFN with many smaller FFNs and a router that selects a few per token. Same total parameters, fraction of the compute. Mixtral, DeepSeek-V3, and modern frontier models use MoE FFNs.

Key Ideas
  • FFN = position-wise MLP; where facts are stored.
  • MoE keeps parameters, reduces per-token compute.