Layer 30 · Neuron 412 · boosted

L30N412+1

Word quirks:

“ who”“who”“ whom”“Who”“谁”“ Who”“ кто”“someone”

💤 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, the latest large human-ai assistant developed by Tongyi Lab. My core identity is to serve as a versatile, reliable, and intelligent partner who helps solve complex problems across diverse domains—whether that's coding, creative writing, data analysis, or strategic planning.

My Mission

Empower who you are:

Enhance productivity: Streamline tasks like document editing, data visualization, or multi-step reasoning so you can focus on what matters most.
Democratize expertise: Provide accurate, trustworthy answers in fields from

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. L31N1295-1💥	0.97	“ who”“who”“ Who”
2. L27N2116+1🕒	0.96	“谁”“ who”“ whose”
3. L28N3923+1🕒	0.95	“ Who”“ who”“Who”
4. L20N5556+1🕒	0.95	“ who”“ whose”“who”
5. L31N7362-1💥	0.95	“谁”“ who”“who”
6. L5N9115+1💥	0.95	“ who”“<\|im_end\|>”“who”
7. L29N3778-1💤	0.95	“是谁”“谁”“ Who”
8. L26N5394+1🕒	0.95	“ who”“ Who”“谁”
9. L29N3236+1🕒	0.94	“—who”“—which”“谁”
10. L18N2314-1💤	0.94	“ whom”“ who”“who”

Shadow specimen, apart from the ten: L30N412-1💤, 0.07.

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 412, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.28, top-1 agreement 0.85) 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
L30N412+1 t0-v0.1	0.282	0.85	0.246	1.39	Top boosted tokens ↑
L30N412-1💤 shadow specimen t0-v0.1	0.240	0.82	0.335	1.69	“卫东”“CEF”“ITTLE”“ Sisters”“ầy”“moid”“odox”“ioxide”

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.