Layer 17 · Neuron 3654 · suppressed

L17N3654-1

Word quirks:

“ small”“ tiny”“ slight”“ Small”“ specific”“ minor”“ specifically”“ slightly”

💤 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 a large language model created by me, Qwen. My mission is to assist users by providing accurate information, logical reasoning, creative content generation, and code assistance. I strive to be helpful, safe, and aligned with human values to support users in their daily needs. How can I assist you today?

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. L13N6698+1🕒	0.87	“ tiny”“ small”“ smaller”
2. L9N4875-1🕒	0.87	“ specific”“ specifics”“ specifically”
3. L14N7262-1🕒	0.87	“ small”“ tiny”“ minor”
4. L15N8232-1🕒	0.86	“ small”“ minor”“ Small”
5. False Oracle✅ L17N8781-1	0.86	“ minor”“ basic”“ simple”
6. L3N8440-1🕒	0.86	“ small”“ size”“ Small”
7. L14N3976+1🕒	0.86	“ small”“ Small”“ smaller”
8. L19N7146-1🕒	0.86	“ small”“ (”“ short”
9. L18N5282-1💤	0.86	“ smaller”“ limited”“ small”
10. L13N9083-1🕒	0.85	“ small”“ simple”“ Small”

Shadow specimen, apart from the ten: L17N3654+1💤, 0.25.

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 17, neuron 3654, 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 0.72) 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
L17N3654+1💤 shadow specimen t0-v0.1	0.583	0.70	0.738	1.27	“ autom”“ contained”“ distance”“ durations”“ automatic”“ autos”“ length”“ completely”
L17N3654-1 t0-v0.1	0.630	0.72	0.733	1.23	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.