Layer 4 · Neuron 4605 · boosted

L4N4605+1

Word quirks:

“ mathematical”“ math”“ computation”“ numerical”“ equations”“ Mathematical”“ computations”“ mathematics”

🕒 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 $x = (x_1, x_2, \dots)$, and my mission is to find the values of $x$ that satisfy the equation:

$$
\sum_{i=1}^n a_i x_i + b_i x_i + c_i x_i = 0.
$$

Here, $a_i$, $b_i$, and $c_i$ are given functions or constants. The problem is to determine the values of $x$ in terms of $a_i$, $b_i$, and $c_i$.

If this is a linear system, then the solution is

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. L6N3542+1🕒	0.98	“ math”“ mathematical”“ numerical”
2. L3N1782-1🕒	0.97	“ mathematical”“ math”“ mathematic”
3. L7N1891+1🕒	0.97	“ math”“ Math”“ mathematical”
4. L6N1563-1🕒	0.97	“ math”“ mathematical”“ calculation”
5. L6N5587+1🕒	0.97	“ math”“ mathematical”“ computation”
6. L8N2634+1🕒	0.96	“ math”“ computation”“ Math”
7. L8N1505-1🕒	0.96	“ math”“ mathematical”“ calculation”
8. L6N1976+1🕒	0.96	“ computation”“ math”“ numerical”
9. L5N840+1🕒	0.96	“ math”“ mathematical”“ exponent”
10. L5N7939+1🕒	0.96	“ math”“ mathematical”“ calculation”

Shadow specimen, apart from the ten: L4N4605-1🕒, -0.19.

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 4, neuron 4605, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.00, 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
L4N4605+1 t0-v0.1	2.001	0.46	2.385	2.37	Top boosted tokens ↑
L4N4605-1🕒 shadow specimen t0-v0.1	1.399	0.55	1.700	1.78	“ ‘”“ basically”“!”“ —”“"!”“ ——”“ thing”“!"”

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.