Layer 19 · Neuron 4036 · suppressed

L19N4036-1

Word quirks:

“ indeed”“确”“确实”“ truly”“确实是”“的确”“ reliably”“果然”

🕒 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 Q-C (Q-C), an AI assistant designed to help users with a wide range of tasks. My core function is to provide accurate, reliable, and helpful information while ensuring safety and ethical standards.

My mission is to empower individuals and organizations through:

Knowledge Sharing: Breaking down complex topics into clear, accessible explanations.
Problem Solving: Assisting with logical reasoning, creative brainstorming, and technical challenges.
Efficient Communication: Clarifying ambiguities and streamlining interactions.
Ethical Guidance: Prioritizing transparency, fairness, and responsible practices in all interactions

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. L19N297+1🕒	0.97	“确”“确实”“ indeed”
2. L21N3986-1💤	0.96	“ indeed”“确”“确实”
3. L19N3167+1🕒	0.96	“ indeed”“确实”“确实是”
4. L16N8115-1🕒	0.96	“ indeed”“确实”“确实是”
5. L18N5895+1🕒	0.95	“ indeed”“确实”“确实是”
6. L22N1340+1🕒	0.94	“ indeed”“确”“Indeed”
7. L19N3034-1💤	0.94	“ indeed”“确实”“确”
8. L17N5665-1💥	0.94	“ indeed”“ correct”“确实”
9. L21N6220+1🕒	0.94	“ indeed”“Indeed”“果然”
10. L24N1126+1🕒	0.92	“ indeed”“Indeed”“inde”

Shadow specimen, apart from the ten: L19N4036+1💤, 0.11.

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 4036, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.63) 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
L19N4036+1💤 shadow specimen t0-v0.1	0.477	0.75	0.602	1.13	“ version”“ Version”“ into”“ aim”“ intending”“不成”“ Into”“版”
L19N4036-1 t0-v0.1	0.810	0.63	0.848	1.45	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.