◄ WORLD II · THE FOLDTHE OCHO · blue builds │ the machine │ red breaks

THE MIXTURE OF EXPERTS

More parameters, same compute. A layer holds N small expert networks and one gate. The gate scores the input, a softmax turns the scores into weights, and only the top-k experts are ever run — the output is the renormalized, gate-weighted sum of just those k. The FLOPs track k, not N: that is how a model buys billions of parameters it does not have to compute. Terms go in, the router fires, the vector comes out. Blue builds it; red breaks it.

source Shazeer et al., Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer (2017) — arxiv.org/abs/1701.06538. Rendered, not quoted.

◧ blue team · builds & defends
3

THE MODEL — gate, route, sum

The layer is four exact steps, no training required to decide them:

1 the gate maps the input to N logits, one per expert. 2 a softmax turns them into a distribution (sums to 1, all ≥ 0). 3 top-k keeps the k highest-weighted experts; the rest are never evaluated. 4 those k weights are renormalized to sum to 1, and the output is their weighted sum of the selected experts' outputs.

Live routing for the current input (fired experts in green):

expertgate pfiredweight
5

THE LINEAGE — scale by sparsity AVAN

A dense transformer block runs one feed-forward network for every token. MoE replaces that single FFN with N of them and a gate — but still runs only k. Parameter count grows ×(N/k) while FLOPs stay flat.

That is the whole downstream story: the dense feed-forward sphere is this one's premise — MoE is what you get when you let the block hold more weight than it spends. The exact neighbouring slug is AMBER pending the FOLD index.

7

THE WITNESS live

The blue team's live check: re-derive the gate softmax, confirm the router fired the k highest gate values, and count that exactly k experts were evaluated. If red reroutes, this badge is where it shows.

▼ the machine ▼
4

DATA IN — the input & the bank in ↓

An input vector x ∈ ℝ⁴ meets a fixed bank of N experts and one gate, all seeded once (deterministic weights). Each expert is an affine map Wᵢx + bᵢ → ℝ³; the gate is a single matrix W_g x → ℝᴺ.

CURRENT INPUT x

Pick an input and a sparsity k below. The gate has never seen a label — it just scores, and the softmax + top-k do the rest.

▼   feed x into the router   ▼
0

▣ THE PANEL — the engine LIT

top-k (highest gate)

Every number below is computed on the spot from the seeded weights — the softmax, the selection, and the weighted sum are never looked up.

LAYER OUTPUT y ∈ ℝ³
▼   the router emits one vector   ▼
8

DATA OUT — the result out ↓

What the machine produces, proven: the gate is a real softmax, top-k selects exactly the k highest-scored experts, the output equals the renormalized gate-weighted sum of only those experts, and the evaluation count is k — not N. That last equality is the sparsity claim.

The blue team's witness (left) recomputes this live; the red team (right) tries to make the wrong experts run.

red team · attacks & breaks ◨
1

THE ADVERSARY

WALL Sparsity is not free. A hard top-k gate is non-differentiable in its choice, so the router trains only through the experts it happens to pick — and it can collapse, sending nearly every token to a few favourites while the rest starve. Shazeer's paper needs an added load-balancing loss and noise just to keep the experts in use.

And the win is memory-bandwidth-bound, not FLOP-bound: the parameters still must be stored and shuffled across devices. "Same compute" is true for the matmuls and quietly false for the interconnect.

2

THE GRAVEYARD

"An MoE with N experts is N× bigger for free." Cut. Capacity grows ×(N/k); the compute saving is real, but routing, replication, and comms overhead are not zero.

"The gate picks the single best expert." Cut. That is only k=1. For k≥2 the output is a renormalized blend of several — the engine computes exactly that sum.

"Unused experts cost nothing." Kept, corrected. They cost no FLOPs, but they cost memory and can go dead — hence the balancing loss the paper adds.

6

THE TAMPER — break it

The red team's move: flip the router to select the bottom-k experts — the lowest gate scores — so the wrong experts run and the output changes. The blue team's witness (window 7) is watching the gate order.

Reroute to the least-favoured experts and the witness recomputes: the fired set is no longer the k highest gate values, so it disagrees with the reference and turns red. Nothing is faked; the attack is real and it is caught.