Layer 10 · Neuron 5209 · suppressed

L10N5209-1

Word quirks:

“ typically”“ usually”“ focus”“!”“ info”“?\”“?”“ broader”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am a large language model. My primary function is to assist with tasks such as problem-solving, data analysis, or completing projects. I can also help generate creative works like stories or poems. How do you like to use me?

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. L12N4046+1💤	0.74	“ info”“ typically”“ usually”
2. L14N7091+1🕒	0.73	“?\”“ typically”“?:”
3. L12N7818+1🕒	0.72	“ usually”“ typically”“ focus”
4. L10N7723-1🕒	0.71	“ typically”“ uncertain”“ clarification”
5. L13N8893-1🕒	0.70	“ info”“ sometimes”“ maybe”
6. L15N5550+1🕒	0.69	“ typically”“ ⏎”“?””
7. L10N4044-1🕒	0.69	“ usually”“ typically”“ ?”
8. L11N689-1🕒	0.68	“ sometimes”“ info”“ usually”
9. L13N3476+1🕒	0.68	“ typically”“ usually”“ typical”
10. L12N2834+1🕒	0.68	“ typically”“ description”“ usually”

Shadow specimen, apart from the ten: L10N5209+1💤, 0.25.

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 10, neuron 5209, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L10N5209+1💤 shadow specimen t0-v0.1	0.948	0.64	1.124	1.31	“ always”“ sometimes”“ everywhere”“ ALWAYS”“ any”“ description”“:"”“ ANY”
L10N5209-1 t0-v0.1	1.332	0.51	1.657	1.53	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.