Layer 19 · Neuron 29 · boosted

L19N29+1

Word quirks:

“av”“ Before”“ Though”“ View”“我们把”“ view”“Before”“学者”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am a large-scale version of the latest model created by the Chinese team at NVIDIA. My name is Qwen, and my mission is to provide users with a strong foundation for the next generation of AI applications. I am designed to help users with various tasks, such as natural language processing, data analysis, and complex problem-solving. I strive to provide accurate, efficient, and ethical assistance to users, ensuring that they have access to the best possible support for their needs. How can I assist you today?
</think>

I am Qwen, a large-scale version of the latest model created by the Chinese team at NVIDIA. My mission

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. L24N1265+1🕒	0.65	“ views”“ View”“view”
2. L30N4373+1💤	0.64	“ view”“视”“ View”
3. L24N3218-1⛔	0.63	“ views”“ viewing”“把”
4. L20N7784-1🕒	0.63	“ view”“-view”“ viewing”
5. L25N1079-1🕒	0.63	“ view”“ viewpoint”“view”
6. L26N588-1🕒	0.63	“观”“ view”“ View”
7. L27N1453-1🕒	0.62	“ view”“ View”“视图”
8. L22N7045-1🕒	0.62	“ views”“ view”“ View”
9. L25N3288-1🕒	0.62	“ view”“View”“ views”
10. L23N8914+1🕒	0.62	“ view”“ View”“ views”

Shadow specimen, apart from the ten: L19N29-1🕒, -0.03.

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 29, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L19N29+1 t0-v0.1	1.130	0.64	1.259	1.43	Top boosted tokens ↑
L19N29-1🕒 shadow specimen t0-v0.1	1.318	0.61	1.419	1.72	“ \”“ texts”“\”“ text”“ textual”“ separate”“ {\”“text”

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.