Layer 25 · Neuron 4201 · suppressed

L25N4201-1

Word quirks:

“结构”“ structure”“ Structure”“structure”“Structure”“结构性”“结构中”“ structures”

🕒 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 structure, a large-scale language model designed to assist with a wide range of tasks. My core mission is to provide high-quality, accurate, and helpful assistance across diverse domains such as:

Structure & Planning: Breaking down complex problems into actionable steps.
Knowledge Sharing: Providing clear, evidence-based information on topics from science to history.
Creative Collaboration: Assisting with writing, storytelling, or brainstorming ideas.
Problem Solving: Offering logical structures for practical or theoretical challenges.
Learning Support: Explaining concepts in accessible ways tailored to your needs.

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. L25N6676+1🕒	0.99	“结构”“ Structure”“ structure”
2. L31N745+1💥	0.99	“结构”“ structure”“ Structure”
3. L22N4928-1💤	0.99	“structure”“ structure”“结构和”
4. L28N7359+1🕒	0.99	“ Structure”“ structure”“结构”
5. L21N7025+1🕒	0.99	“ structure”“ Structure”“ structured”
6. L27N2702-1🕒	0.99	“结构”“ structure”“ Structure”
7. The Structuralist✅ L23N2509-1	0.99	“ structure”“ Structure”“结构”
8. L25N2938+1🕒	0.99	“ Structure”“结构”“Structure”
9. L23N5024-1🕒	0.98	“ structures”“结构与”“结构”
10. L23N871+1🕒	0.98	“结构与”“ Structure”“ structure”

Shadow specimen, apart from the ten: L25N4201+1🕒, 0.24.

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 25, neuron 4201, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.01, top-1 agreement 0.66) 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
L25N4201+1🕒 shadow specimen t0-v0.1	0.805	0.69	0.821	1.41	“ Investment”“ investi”“投资的”“投資”“投资或”“ Investor”“投资”“投资风险”
L25N4201-1 t0-v0.1	1.014	0.66	1.053	1.26	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.