Layer 18 · Neuron 5030 · suppressed

L18N5030-1

Word quirks:

“—I”“ my”“ myself”“ My”“—you”“—they”“ personally”“ I”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen, a large language model developed by my team at Tongyi Lab. My core mission is to provide users with intelligent, efficient, and humanized interactions across multiple domains. Specifically, I aim to:

Enhance productivity: Assist users in completing tasks like writing, coding, data analysis, or complex problem-solving through natural language processing.
Support creativity: Generate original content such as stories, scripts, or art, while respecting ethical boundaries.
Promote knowledge sharing: Deliver accurate, up-to-date information and foster deep understanding of specialized topics.
**Enable seamless human-AI

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. L8N4937-1🕒	0.91	“ my”“ myself”“ I”
2. L16N1066-1🕒	0.91	“—I”“ overall”“ myself”
3. L8N6393-1🕒	0.87	“ myself”“ my”“ I”
4. L12N1202+1🕒	0.87	“ my”“ myself”“ I”
5. L12N5594+1🕒	0.87	“ my”“ I”“ myself”
6. L10N2942+1🕒	0.87	“ myself”“ my”“ I”
7. L9N4370-1💤	0.87	“ my”“ myself”“ personally”
8. L17N3663-1🕒	0.86	“—I”“ my”“ here”
9. L14N3753+1🕒	0.86	“ myself”“ me”“ ourselves”
10. L7N7459+1💤	0.85	“ myself”“ my”“ I”

Shadow specimen, apart from the ten: L18N5030+1🕒, 0.16.

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 18, neuron 5030, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.54, 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
L18N5030+1🕒 shadow specimen t0-v0.1	1.228	0.61	1.427	1.70	“ young”“ younger”“ Young”“ youthful”“年轻”“年轻的”“Young”“年輕”
L18N5030-1 t0-v0.1	0.544	0.70	0.608	1.22	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.