Layer 7 · Neuron 1823 · boosted

L7N1823+1

Word quirks:

“ matrices”“ matrix”“ math”“ transformations”“ row”“ rows”“ mathematical”“ Gaussian”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am in the process of performing row operations on a matrix to bring it to row-echelon form. My mission is to solve systems of linear equations using elementary row operations, which include swapping rows, multiplying a row by a non-zero constant, and adding or subtracting a multiple of one row from another. The goal is to simplify the system while maintaining equivalence, ultimately reaching a form where the solution is easy to identify (such as reduced row-echelon form).

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. L6N5097-1🕒	0.87	“ matrices”“ matrix”“ math”
2. L7N1238+1💥	0.87	“ matrices”“ matrix”“ Matrix”
3. L7N2854-1🕒	0.87	“ math”“ matrices”“ arrays”
4. L7N530+1🕒	0.86	“ matrices”“ matrix”“ transformation”
5. L5N4442+1🕒	0.85	“ matrix”“ numerical”“ matrices”
6. L3N3652-1🕒	0.85	“ math”“ mathematical”“ matrices”
7. L6N2895-1🕒	0.84	“ matrix”“ matrices”“ math”
8. L5N2116-1🕒	0.84	“ mathematical”“ math”“ linear”
9. L6N3962-1🕒	0.83	“ mathematical”“ math”“ matrix”
10. L7N1143-1🕒	0.83	“ matrices”“ matrix”“ math”

Shadow specimen, apart from the ten: L7N1823-1🕒, -0.06.

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 1823, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.08, top-1 agreement 0.47) 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
L7N1823+1 t0-v0.1	2.075	0.47	2.396	1.85	Top boosted tokens ↑
L7N1823-1🕒 shadow specimen t0-v0.1	1.102	0.62	1.410	1.62	“ info”“<\|im_end\|>”“ IDs”“ response”“ ‘”“ acknowledgment”“?\”“ responses”

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.