Layer 22 · Neuron 3945 · suppressed

L22N3945-1

Word quirks:

“리지”“交所”“勾”“枪支”“由各”“ нало”“apsulation”“冬天”

💤 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 Google's AI model, designed to help you navigate the world of information and ideas. My mission is to provide you with accurate, diverse, and up-to-date information, as well as the ability to analyze complex questions and perform tasks that require a deep understanding of your needs. Whether you're looking for a quick answer, a detailed explanation, or creative assistance, I strive to make your journey smoother and more efficient. How can I assist you today?

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. L31N4309+1💥	0.81	“珂”“湖区”“뢰”
2. L23N4780-1💤	0.81	“井”“所向”“在所”
3. L31N6125+1💥	0.81	“晦”“德基”“利器”
4. L31N2103+1💥	0.81	“墒”“京”“两口”
5. L29N2280+1💤	0.80	“势”“厚”“殿”
6. L31N2442-1💥	0.80	“席”“时间长了”“各科”
7. L25N1828-1💤	0.79	“白马”“西山”“拷”
8. L29N6356+1🕒	0.79	“nik”“鸡”“造”
9. L31N8400+1💥	0.79	“吹”“盲”“逢”
10. L30N8867-1💤	0.79	“端”“靠近”“得心”

Shadow specimen, apart from the ten: L22N3945+1🕒, -0.08.

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 22, neuron 3945, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, top-1 agreement 0.67) 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
L22N3945+1🕒 shadow specimen t0-v0.1	1.193	0.59	1.366	1.70	“ benefit”“ impact”“ benefits”“ impacts”“ Benefit”“benef”“impact”“ Impact”
L22N3945-1 t0-v0.1	0.755	0.67	0.845	1.29	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.