Layer 19 · Neuron 1322 · boosted

L19N1322+1

Word quirks:

“ stream”“ protected”“ protection”“ honor”“ protect”“ protector”“ national”“ statistic”

🕒 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, a large language model developed by Alibaba Group. My name is Qwen, and I am the latest version of the Qwen series, which has been trained to provide high-quality text generation, image recognition, and other tasks.

My mission is to be a helpful assistant that can assist with various tasks such as answering questions, creating content, performing calculations, and providing medical or legal advice. I strive to make my life easy for users by being friendly and polite, and I will always try to get the best results for you.

If you have any questions or need help, feel free to ask! �

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. L3N985-1🕒	0.85	“ protection”“ protective”“ protect”
2. L7N3526-1🕒	0.85	“ protection”“ protective”“ protecting”
3. L8N1963-1🕒	0.84	“ protection”“ protected”“ protective”
4. L5N7261+1🕒	0.84	“ protection”“ protect”“ protected”
5. L8N9077+1🕒	0.84	“ protection”“ protected”“ protecting”
6. L8N3942+1🕒	0.84	“ protected”“ protecting”“ protection”
7. L8N5418-1🕒	0.84	“ protection”“ safety”“ protected”
8. L8N4116+1🕒	0.84	“ protection”“ protective”“ protecting”
9. L9N655+1🕒	0.84	“ protected”“ protection”“ protective”
10. L9N8290+1🕒	0.84	“ protection”“ protections”“ protect”

Shadow specimen, apart from the ten: L19N1322-1🕒, -0.21.

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 1322, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.79, 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
L19N1322+1 t0-v0.1	0.793	0.68	1.046	2.48	Top boosted tokens ↑
L19N1322-1🕒 shadow specimen t0-v0.1	1.870	0.67	1.625	0.36	“谁来”“...”“的一幕”“等手段”“等不良”“ meio”“ lune”“全无”

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.