Layer 10 · Neuron 6189 · boosted

L10N6189+1

Word quirks:

“ *"”“ "*”“ describing”“ *”“ "**”“:*”“ *“”“ "`”

💤 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 Qwen, a large-scale language model developed by the Tongyi Lab (part of Alibaba Group). My name is Qwen3.5, and I am designed to assist users with complex tasks across a wide range of domains.

My mission is to provide high-quality, accurate, and helpful information, code generation, and logical reasoning to users around the globe. Specifically:

Capabilities: I support fluent text generation in over 100 languages and can handle complex reasoning and analysis.
Applications: I can help with tasks like writing code, analyzing data, creating content, and

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. L13N376+1💤	0.94	“ "”“ "”“ *”
2. L16N8461+1🕒	0.94	“ "”“ basic”“ "”
3. L8N6137+1💤	0.93	“ ”“ “”“ *"”
4. L11N1935-1🕒	0.93	“ "”“ "”“ *“”
5. L15N7952+1🕒	0.93	“ "”“ "”“ *«”
6. L13N6189-1💤	0.93	“ “”“ "”“ *”
7. L12N403-1🕒	0.93	“ ”“ "”“ *“”
8. L17N1956-1🕒	0.92	“ "”“ ”“ *"”
9. L12N2725+1🕒	0.92	“ “”“ "”“ **“”
10. L13N7040-1🕒	0.92	“ ”“ "”“ "*”

Shadow specimen, apart from the ten: L10N6189-1🕒, -0.17.

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 10, neuron 6189, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.91, 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
L10N6189+1 t0-v0.1	0.906	0.59	1.097	1.29	Top boosted tokens ↑
L10N6189-1🕒 shadow specimen t0-v0.1	1.430	0.55	1.748	2.27	“ ONLY”“ ‘”“ special”“ corresponding”“ specially”“ judgement”“ travelling”“ replacement”

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.