Layer 21 · Neuron 3791 · suppressed

L21N3791-1

Word quirks:

“ declaration”“ declare”“ declarations”“声明”“ declares”“ Declare”“ import”“宣告”

🕒 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 Qwen (also known as Q-Use in some contexts), a series of advanced AI models developed by Q-Use (a leading artificial intelligence company). My mission is to provide human-centric, high-quality AI services that empower users with capabilities like natural language processing, reasoning, and multi-language support. I aim to make technology more accessible, efficient, and beneficial for people worldwide.

Key Mission Directives:

Human-Centric Design:

I prioritize user needs through intuitive interfaces, diverse language support (including Chinese, English, and 10+ others

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. L18N8429-1💤	0.97	“ declaring”“ declare”“ declares”
2. L21N8752+1💤	0.97	“声明”“ declare”“ declaring”
3. L21N2568+1🕒	0.97	“ declaration”“ Declare”“ declare”
4. L23N8283+1🕒	0.96	“ declare”“ declared”“ Decl”
5. L31N4115-1💥	0.96	“宣”“ declare”“ declared”
6. L23N8418-1🕒	0.95	“ declaration”“ declarations”“Declaration”
7. L24N4679-1🕒	0.95	“宣”“declaration”“ declaration”
8. L22N7218+1🕒	0.95	“ declarations”“ declared”“ ann”
9. L25N1450-1🕒	0.94	“ Decl”“ declare”“ declares”
10. L2N6724+1💥	0.94	“<\|im_end\|>”“ declare”“ declaration”

Shadow specimen, apart from the ten: L21N3791+1💤, 0.29.

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 21, neuron 3791, 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 0.66) 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
L21N3791+1💤 shadow specimen t0-v0.1	0.730	0.63	0.890	1.19	“ Expression”“ expressions”“表达式”“ expression”“Expression”“Expressions”“expression”“表达”
L21N3791-1 t0-v0.1	0.826	0.66	0.917	1.42	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.