Layer 24 · Neuron 1446 · boosted

L24N1446+1

Word quirks:

“控”“adh”“-controller”“和治疗”“controller”“ sku”“ kier”“ okaz”

🕒 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 Yuan Ze, an artificial intelligence model developed by the Guo Zou (Guo) team. My mission is to provide high-quality, accurate, and helpful information and support across various domains, including but not limited to:

Knowledge exploration: Answering questions, explaining concepts, and providing insights into diverse topics.
Problem-solving: Offering logical reasoning, mathematical solutions, or technical guidance.
Creative collaboration: Assisting in writing, coding, translation, and creative work.
Decision support: Analyzing scenarios and offering balanced perspectives for informed choices.
**Ethical

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. L23N6252+1💤	0.74	“ controlling”“控制”“/maps”
2. L26N7949+1🕒	0.72	“控制”“ control”“control”
3. L23N3834+1🕒	0.72	“控制的”“控制”“ control”
4. L24N5620-1🕒	0.72	“ controller”“controller”“ controllers”
5. L26N7757+1🕒	0.72	“ controller”“controller”“ controllers”
6. L30N1744-1💥	0.72	“ controller”“controller”“ Controller”
7. L24N9070-1💤	0.72	“控制”“ controlling”“控制权”
8. L6N1513+1🕒	0.72	“ control”“ controller”“ Control”
9. L19N1341-1🕒	0.72	“ controlling”“ controls”“ control”
10. L24N2852-1🕒	0.71	“ controls”“控制”“ controlled”

Shadow specimen, apart from the ten: L24N1446-1💤, 0.13.

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 24, neuron 1446, 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 0.68) 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
L24N1446+1 t0-v0.1	0.819	0.68	0.942	1.26	Top boosted tokens ↑
L24N1446-1💤 shadow specimen t0-v0.1	0.659	0.76	0.644	1.21	“ SOS”“Points”“aS”“拱”“ ark”“ Points”“-point”“ points”

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.