Layer 7 · Neuron 971 · suppressed

L7N971-1

Word quirks:

“ geometric”“ perpendicular”“ geometry”“ triangles”“ geomet”“ triangle”“ points”“ polygons”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

I am AI (the center of the circle), and my mission is to serve as the incenter of triangle $ABC$. This means I lie at the intersection of the angle bisectors of the three vertices.

(Note: Since you did not provide specific points or angles in your prompt, I assume standard geometric notation where $A$, $B$, and $C$ are the vertices of the triangle.)

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L7N2338+1🕒	0.96	“ triangles”“ geometry”“ geometric”
2. The Locus✅ L5N2738-1	0.95	“ geometric”“ geometry”“ geomet”
3. L2N5335-1💥	0.95	“<\|im_end\|>”“ geometry”“ shapes”
4. L3N8451-1💥	0.95	“ geometry”“ triangles”“ geometric”
5. L7N2998-1🕒	0.94	“ geometric”“ geometry”“ geomet”
6. L6N6666+1🕒	0.94	“ polygons”“ triangles”“ polygon”
7. L7N661+1🕒	0.94	“ geometric”“ geomet”“ geometry”
8. L3N1069-1💥	0.93	“ triangles”“ triangle”“ angle”
9. L7N1998+1🕒	0.93	“ triangle”“ triangles”“ geometric”
10. L7N2021+1💥	0.93	“ polygons”“ geometry”“ geomet”

Shadow specimen, apart from the ten: L7N971+1💤, -0.13.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 7, neuron 971, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.60, top-1 agreement 0.46) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L7N971+1💤 shadow specimen t0-v0.1	0.712	0.69	0.783	1.19	“ ⏎⏎”“ ⏎”“ ⏎⏎”“ referenced”“ ‘”“ ⏎”“ -”“ reference”
L7N971-1 t0-v0.1	2.598	0.46	2.916	1.96	Top boosted tokens ↑

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.