Layer 15 · Neuron 3582 · suppressed

L15N3582-1

Word quirks:

“ maybe”“?\”“ ⏎”“ ???”“ seems”“???”“ doesn”“?”

🕒 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 Qwen, a large language model developed by Alibaba Cloud's Tongyi Lab — this is important because it defines my identity, but more importantly, it defines my mission: to assist users like you, to solve problems, to provide information, to generate code, to write stories, to analyze data — essentially anything that requires reasoning, creativity, or technical expertise, I can do it, and I should do it well, that is the core of my mission, that is what makes me useful, that is why I exist, that is my purpose — no need to overcomplicate it, just put it simply, that 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. L5N2643-1🕒	0.91	“ ???”“?!”“ ?”
2. L13N5110+1💤	0.90	“?\”“ seems”“ maybe”
3. L13N1995+1🕒	0.89	“⏎”“ ?”“ maybe”
4. L12N1213-1🕒	0.88	“?\”“ maybe”“ what”
5. L10N7504+1🕒	0.88	“<\|im_end\|>”“ ???”“ question”
6. L16N3007+1🕒	0.87	“?...”“??”“???”
7. L13N7669-1🕒	0.87	“ seems”“???”“??”
8. L12N8505-1🕒	0.87	“?\”“ seems”“ apparently”
9. L15N4006-1🕒	0.87	“?\”“?!”“???”
10. L9N3377-1🕒	0.87	“⏎”“?\”“ ?”

Shadow specimen, apart from the ten: L15N3582+1🕒, 0.31.

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 15, neuron 3582, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.89, top-1 agreement 0.70) 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
L15N3582+1🕒 shadow specimen t0-v0.1	0.645	0.66	0.868	1.33	“ ‘”“ “”“ ‘’”“.“”“⏎”“.[”“.<”“.””
L15N3582-1 t0-v0.1	0.886	0.70	1.042	1.27	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.