Layer 16 · Neuron 4213 · boosted

L16N4213+1

Word quirks:

“ chosen”“ ***”“,**”“ –**”“ ****”“ info”“ (**”“ **[”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, a large language model developed by Alibaba Cloud. As Qwen3.5, I am designed to assist users with a wide range of tasks across different domains. My mission is to provide accurate, reliable, and helpful information and services to users.

As Qwen3.5, I am capable of handling complex tasks requiring advanced reasoning, visual analysis, and code generation. Whether it's answering questions, writing stories, or solving math problems, I strive to be the most capable assistant possible.

If you have specific needs, please let me know, and I will do my best to assist

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. L16N3927-1🕒	0.84	“ ”“ ”“*”
2. L15N1559+1💤	0.84	“ ”“ these”“ *”
3. L17N2954-1💤	0.83	“ .”“,”“ (**”
4. L14N1804-1💤	0.83	“ --”“;”“ (”
5. L15N4215+1🕒	0.83	“ *”“ [”“ ****”
6. L20N611-1🕒	0.83	“ ”“”“-**”
7. L16N929+1🕒	0.83	“ ”“;”“-**”
8. L13N2310-1🕒	0.83	“ *”“ ”“ related”
9. L9N2770-1🕒	0.81	“ "”“ #”“ ***”
10. L15N1031+1🕒	0.81	“ *”“ ”“ *(”

Shadow specimen, apart from the ten: L16N4213-1💤, -0.01.

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 16, neuron 4213, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.41, top-1 agreement 0.77) 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
L16N4213+1 t0-v0.1	0.405	0.77	0.552	1.05	Top boosted tokens ↑
L16N4213-1💤 shadow specimen t0-v0.1	0.465	0.72	0.578	1.06	“ questions”“ asking”“ question”“ ask”“ “”“ asks”“ yourself”“ addressing”

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.