Layer 18 · Neuron 4251 · boosted

L18N4251+1

Word quirks:

“ measure”“ measuring”“ Measure”“measure”“衡量”“ measurement”“ measures”“ metrics”

🕒 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 Qun, a specialized AI capability designed for the People's Bank of China (PBC) and its designated compliance officers. As a regulated financial domain, I focus on delivering secure, standardized, and auditable outputs—ensuring that every metric, risk indicator, or regulatory report meets strict compliance standards.

My mission is to provide a unified, high-impact interface for financial performance tracking, enabling stakeholders to measure progress across key areas such as customer experience, operational efficiency, and strategic outcomes. By aligning with industry-standard metrics, I ensure that all deliverables are both meaningful and actionable.

If you'd like to

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. L13N7107-1🕒	0.98	“ measure”“ measuring”“衡量”
2. L26N1391+1🕒	0.98	“ measuring”“测量”“ measurement”
3. L23N3213+1🕒	0.97	“ measuring”“测量”“ measure”
4. L2N3308-1💥	0.97	“ measurement”“ measure”“ measuring”
5. L23N5670-1🕒	0.97	“ measuring”“测量”“ measurements”
6. L24N6962-1🕒	0.97	“.measure”“ Measure”“测量”
7. L21N1859-1💤	0.96	“_measure”“Measure”“ measure”
8. L24N7896-1🕒	0.96	“ measure”“ Measure”“.measure”
9. L1N3172-1💥	0.96	“ measure”“ measurement”“measure”
10. L21N284-1🕒	0.96	“ measure”“ measured”“ measures”

Shadow specimen, apart from the ten: L18N4251-1💤, 0.02.

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 18, neuron 4251, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.97, top-1 agreement 0.59) 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
L18N4251+1 t0-v0.1	0.974	0.59	1.065	1.30	Top boosted tokens ↑
L18N4251-1💤 shadow specimen t0-v0.1	0.606	0.69	0.673	1.28	“ just”“ handling”“ excluding”“ based”“ generally”“ basically”“ without”“ rather”

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.