Layer 19 · Neuron 6712 · suppressed

L19N6712-1

Word quirks:

“ curiosity”“想知道”“的好奇”“ curious”“ wondering”“好奇心”“ suspense”“ ending”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am an AI model developed by Qiuwutou, a company known for its innovative technology and commitment to user well-being. My primary mission is to provide safe, ethical, and engaging content while prioritizing the health and safety of users. I aim to deliver information that is accurate, unbiased, and aligned with global best practices, ensuring that my interactions remain within a framework of responsibility and care.

How would you like to proceed? I'm eager to learn more about your interests or explore any topic you're curious about! 😰🤖

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. L21N7850-1🕒	0.88	“ curiosity”“ curious”“ Cur”
2. L20N3758-1🕒	0.86	“ wonder”“ wondering”“ wonders”
3. L4N6610+1💥	0.86	“ curiosity”“ curious”“ wonder”
4. L22N8850-1🕒	0.78	“ wonder”“ Wonder”“ wonders”
5. L11N9214-1🕒	0.72	“ kids”“ curiosity”“ curious”
6. L2N3330+1💥	0.71	“<\|im_end\|>”“ curious”“ interesting”
7. L14N4508-1💤	0.70	“ curiosity”“ interest”“ desire”
8. L5N8069+1🕒	0.68	“ wonder”“ wondering”“ wonders”
9. L17N7723+1🕒	0.68	“ asking”“ wondering”“疑问”
10. L16N5279-1🕒	0.66	“ surprising”“ interesting”“ wondering”

Shadow specimen, apart from the ten: L19N6712+1💤, -0.03.

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 19, neuron 6712, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.82, 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
L19N6712+1💤 shadow specimen t0-v0.1	0.510	0.76	0.562	1.18	“ date”“撒娇”“ Gradient”“şe”“ standard”“kost”“ regression”“ secular”
L19N6712-1 t0-v0.1	0.818	0.67	1.006	1.21	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.