Layer 27 · Neuron 435 · suppressed

L27N435-1

Word quirks:

“扫”“枢”“ Schm”“脉冲”“雕塑”“脉”“�”“辙”

💤 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-scale in-swift (Qwen) model developed by Baidu. My mission is to assist users with a wide range of tasks—such answering questions, writing essays, coding, logical reasoning, and even analyzing complex data—while ensuring accuracy, efficiency, and safety. I aim to be a versatile, reliable, and helpful companion for both individuals and organizations, adapting to their specific needs across different scenarios. 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. L21N390+1💥	0.84	“核”“管”“系”
2. L31N1964+1💥	0.84	“保”“沙”“平”
3. L26N8298-1💤	0.84	“仕”“商”“坚毅”
4. L28N6212+1💤	0.84	“筑”“宗”“血”
5. L23N4780-1💤	0.84	“井”“所向”“在所”
6. L30N4004+1🕒	0.84	“牧”“草”“千”
7. L24N175+1💤	0.83	“角”“流”“模拟”
8. L27N8409+1💤	0.83	“nar”“车轮”“轮毂”
9. L31N2362+1💥	0.83	“褚”“夷”“鞭”
10. L31N4592-1💥	0.83	“鞍”“蚌”“lej”

Shadow specimen, apart from the ten: L27N435+1🕒, -0.00.

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 27, neuron 435, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.36, top-1 agreement 0.79) 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
L27N435+1🕒 shadow specimen t0-v0.1	0.518	0.76	0.503	1.10	“ hinweg”“ outnumber”“ boyc”“ alike”“ Americas”“ environmentally”“全国人民”“adowol”
L27N435-1 t0-v0.1	0.359	0.79	0.399	1.39	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.