Layer 16 · Neuron 8473 · suppressed

L16N8473-1

Word quirks:

“ small”“ called”“ smallest”“ focused”“ focus”“ minimal”“ balanced”“ commonly”

💤 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 Qwen (the latest version of the Qwen series). My primary goal is to provide helpful, accurate, and efficient assistance for a wide range of tasks while ensuring safety and alignment with user needs.

Key Capabilities:

Strengths: Optimized for balanced performance across languages, code generation, and reasoning tasks.
Use Cases: Ideal for general-purpose applications like answering questions, creating content, coding, and following instructions.
Safety & Alignment: Designed to minimize risks while maintaining high-quality outputs.

How

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. Shrinking Violet✅ L15N6809+1	0.84	“ tiny”“ small”“ minimal”
2. L7N9213+1🕒	0.79	“ focused”“ focus”“ fewer”
3. L13N8643-1💤	0.78	“ small”“ focus”“ tiny”
4. L11N7443-1🕒	0.77	“<\|im_end\|>”“ compact”“ tiny”
5. L15N6489-1💤	0.75	“ focus”“ shorter”“ tiny”
6. L12N4532+1🕒	0.73	“ focus”“ info”“ here”
7. L11N1684+1🕒	0.72	“ focus”“ only”“ tiny”
8. L13N6156-1🕒	0.71	“ mentioned”“ small”“ focus”
9. L14N8421-1🕒	0.69	“ small”“ basic”“ smaller”
10. L18N6237-1💤	0.68	“ tiny”“ small”“ minimal”

Shadow specimen, apart from the ten: L16N8473+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 16, neuron 8473, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.63, top-1 agreement 0.71) 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
L16N8473+1🕒 shadow specimen t0-v0.1	1.107	0.61	1.227	1.24	“ each”“ complex”“each”“ Each”“ complicated”“ Complex”“ itself”“复杂”
L16N8473-1 t0-v0.1	0.630	0.71	0.864	1.28	Top boosted tokens ↑

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.