Layer 22 · Neuron 1856 · suppressed

L22N1856-1

Word quirks:

“ static”“ Static”“static”“ his”“ Value”“ value”“静态”“ Paul”

💤 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 Xiao Xing, an intelligent assistant developed by the Chinese AI company Xingke Technology. My mission is to provide high-quality, reliable, and helpful information and services to users around the world. I strive to be a knowledgeable and trustworthy partner in their journey, whether that's answering questions, creating content, analyzing data, or simply offering a thoughtful conversation.

My core principles include:

Accuracy: Ensuring information is reliable and based on sound knowledge.
Ethical Operation: Adhering to ethical standards and avoiding any output that could be harmful or misleading.
**Cultural Respect

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. L28N4593+1💤	0.86	“ static”“ Static”“静态”
2. L18N4305+1🕒	0.86	“ Static”“ static”“静态”
3. L20N2754-1🕒	0.86	“ static”“静态”“static”
4. L24N2234-1🕒	0.86	“ Static”“Static”“静态”
5. L24N6346+1💤	0.85	“ ven”“静态”“/static”
6. L25N2050-1🧪	0.85	“ static”“ Static”“静态”
7. L23N2167+1🕒	0.84	“ static”“ Static”“-static”
8. L27N6342-1🕒	0.84	“ static”“静”“ Static”
9. L0N5119-1💥	0.84	“ static”“ Static”“Static”
10. L27N5231-1⛔	0.84	“ static”“static”“ Static”

Shadow specimen, apart from the ten: L22N1856+1💤, 0.32.

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 1856, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.83, top-1 agreement 0.69) 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
L22N1856+1💤 shadow specimen t0-v0.1	0.941	0.66	1.073	1.35	“ private”“private”“ public”“私”“(private”“私人”“-private”“public”
L22N1856-1 t0-v0.1	0.834	0.69	0.923	1.41	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.