Layer 30 · Neuron 3051 · boosted

L30N3051+1

Word quirks:

“aczy”“네요”“jut”“eceği”“gó”“ará”“ungsi”“utura”

💤 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 Qwen3.5, a large language model developed by Tongyi Lab (part of Alibaba Group). My primary mission is to assist users in solving complex problems, generating creative content, and providing actionable insights across a wide range of domains—while ensuring safety, accuracy, and adaptability to diverse contexts.

Key aspects of my mission:

Practical Utility:

Help with writing, coding, data analysis, and logical reasoning to streamline workflows and boost efficiency.
Offer real-time information retrieval (via tools like search) and **long

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. L27N365+1💥	0.69	“utre”“chte”“선”
2. L31N472+1💥	0.67	“êu”“гами”“iedy”
3. L31N7838+1💥	0.67	“agar”“culp”“ naro”
4. L30N1736+1🕒	0.67	“jus”“uzzi”“adius”
5. L26N3526+1💤	0.67	“atsu”“acz”“ùm”
6. L30N2063+1💤	0.67	“utch”“iau”“łady”
7. L30N4938-1💤	0.67	“uger”“ustain”“ inn”
8. L31N2455+1💥	0.66	“�”“uchos”“ottom”
9. L31N5911+1💥	0.66	“urre”“otov”“赫”
10. L31N2718+1💥	0.66	“uret”“agar”“ubu”

Shadow specimen, apart from the ten: L30N3051-1💤, 0.37.

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 30, neuron 3051, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.25, top-1 agreement 0.84) 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
L30N3051+1 t0-v0.1	0.247	0.84	0.292	1.62	Top boosted tokens ↑
L30N3051-1💤 shadow specimen t0-v0.1	0.246	0.88	0.272	1.42	“ecer”“moc”“ECO”“slag”“elf”“外商”“ЕК”“кро”

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.