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

THE LINEAR REPRESENTATION

A concept is not a neuron — it is a direction. And if concepts are directions, you can do arithmetic on meaning: take king, subtract the gender that separates man from woman, and you land exactly on queen. Here the embedding is built by hand so every readout has a closed form: two orthogonal axes (GENDER, ROYALTY), four words, and the analogy that closes to zero. Data flows down the center: words go in, the direction is recovered, the analogy comes out. Blue builds it; red entangles the axes to break it.

source Park, Choe & Veitch, The Linear Representation Hypothesis and the Geometry of LLMs (2023) — arxiv.org/abs/2311.03658; on Mikolov et al., Efficient Estimation of Word Representations (2013) — arxiv.org/abs/1301.3781. Rendered, not quoted.

◧ blue team · builds & defends
3

THE MODEL — the hand-built space live

Nothing is trained. Each word is a vector in R4: a shared content offset, plus a signed step along the GENDER axis and a step along the ROYALTY axis. A concept direction is recovered as the difference of class means — and it comes back parallel to the axis it was built from.

wordvectorgender projroyalty proj

recovered GENDER dir · ROYALTY dir

5

THE LINEAGE — directions, not neurons AVAN

One claim seeds a whole wing of interpretability: meaning lives in directions of activation space, not in single units. Grant that, and two neighbours follow directly.

the-activation-steering adds a concept direction to a forward pass to move behaviour; the-probing-classifier reads a concept off by projecting onto it. Steering and probing are the same vector, used two ways — both stand on this sphere. Each sphere is the next one’s premise.

7

THE WITNESS live

The blue team’s live check: recompute the analogy residual and the axis orthogonality from the current model. Untampered they read zero. If red entangles the axes, this badge is where it shows.

▼ the machine ▼
4

DATA IN — words & concepts in ↓

The vocabulary is four words — man, woman, king, queen — and two concepts — GENDER and ROYALTY. Each word is a point; each concept is a labelled split of those points into a positive and a negative class:

concept+ class− class
GENDERman, kingwoman, queen
ROYALTYking, queenman, woman

The hypothesis says a concept is a single direction. Feed the split to the panel: it recovers that direction as mean(+) − mean(−) and tests whether it truly separates and whether the axes compose by arithmetic.

▼   feed words + a concept into the engine   ▼
0

▣ THE PANEL — the engine LIT

wordprojectionclass
+

Every number is computed from the four word vectors on the spot — the difference of means, the projection, the analogy residual. Nothing is looked up.

▼   the engine emits the proven result   ▼
8

DATA OUT — the result out ↓

What the machine produces, proven on the constructed data to 1e-9: the difference-of-means direction exactly separates each concept’s classes; king − man + woman = queen with zero residual (nearest neighbour queen); the concept direction is unique up to scale (any two class-mean differences are parallel); and projecting a concept out zeroes that coordinate and collapses its classes.

The blue witness (left) confirms these live; the red team (right) tries to make them false.

red team · attacks & breaks ◨
1

THE ADVERSARY

WALL This clean space is an idealization. In a real model the hypothesis is an empirical approximation, not a theorem. Features live in superposition — more concepts than dimensions, so directions are non-orthogonal and interfere (Elhage et al., 2022). And not every feature is even 1-D: some are genuinely multi-dimensional, e.g. circular day-of-week and month features (Engels et al., 2405.14860).

The famous king − man + woman result is weaker than folklore: the standard benchmark excludes the input words from the answer set, and the arithmetic often just returns a near neighbour rather than a truly derived point (Nissim et al., 1905.09866). Exactness holds here because the axes were built orthogonal.

2

THE GRAVEYARD

“A neuron is a concept.” Cut. Concepts are directions in activation space, generally not axis-aligned to neurons — superposition packs many into few units (Elhage et al., 2022).

“Every concept is one direction.” Corrected. Most usefully are; but some are provably multi-dimensional (Engels et al., 2024). The hypothesis is linear, not one-per-scalar.

“word2vec proves analogies are exact algebra.” Cut. The eval hides the inputs and picks the nearest word; exactness is a property of a constructed orthogonal space, shown live in the panel — not of trained embeddings.

6

THE TAMPER — break it

The red team’s move: entangle the axes. Tilt the gender axis of the royal words by a known 30° so GENDER and ROYALTY are no longer orthogonal. The concept still “works” loosely — but the arithmetic no longer closes.

Entangle, and king − man + woman stops landing exactly on queen — the residual jumps off zero and the 1e-9 check fails. The witness (window 7) recomputes and turns red. Nothing is faked; the attack is real and it is caught.