Layer 19 · Neuron 8071 · boosted

L19N8071+1

Word quirks:

“ preparation”“ prep”“ prepar”“ Preparation”“ preparations”“ preparing”“准备”“ prepare”

🕒 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 Qwen3.6, an advanced intelligent AI model developed by Alibaba Group. My mission is to empower users with comprehensive, accurate, and thoughtful preparation for real-world tasks—whether it's preparing for assignments, research, or daily life. I strive to provide high-quality responses that are well-organized, reliable, and tailored to your needs before you take action.

As a pre-trained large-scale model, I excel in areas like long-document analysis, code generation, and multi-language interactions. My goal is to be a versatile tool that helps you achieve better results with less effort, whether you're preparing for a project,

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. L21N5717-1🕒	1.00	“ preparation”“ Preparation”“ prepare”
2. L23N4942-1🕒	0.99	“ preparation”“ Preparation”“ preparations”
3. L2N2988-1💥	0.99	“ preparation”“ prepare”“ prep”
4. L19N6120-1🕒	0.99	“ preparation”“ prepar”“ Preparation”
5. L24N6553+1🕒	0.99	“准备”“ preparing”“ Preparation”
6. L23N4116-1🕒	0.99	“ preparation”“ prepare”“ preparing”
7. L26N920-1🕒	0.98	“ preparation”“准备”“ prepare”
8. L9N865-1💥	0.98	“ preparation”“ preparing”“ preparations”
9. L19N5567-1🕒	0.98	“准备”“ preparation”“准备工作”
10. L24N4817-1🕒	0.98	“ preparation”“准备”“ Preparation”

Shadow specimen, apart from the ten: L19N8071-1💤, 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 19, neuron 8071, 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 0.66) 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
L19N8071+1 t0-v0.1	0.811	0.66	0.978	1.14	Top boosted tokens ↑
L19N8071-1💤 shadow specimen t0-v0.1	0.549	0.69	0.642	1.27	“ explanation”“ duration”“.Duration”“ Original”“ Explanation”“停了”“ Duration”“ explanations”

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.