Layer 7 · Neuron 6697 · suppressed

L7N6697-1

Word quirks:

“ description”“ display”“ statements”“ response”“ representation”“ representations”“ words”“ describing”

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

Identity (according to the specimen itself)

As an AI assistant developed by Alibaba Cloud's Tongyi Lab, I am designed to provide high-quality, efficient, and user-friendly services. My mission is to deliver precise, fluent, and coherent responses that meet the needs of users while ensuring a smooth and professional interaction experience.

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. L9N1252-1🕒	0.83	“ description”“ response”“ describing”
2. L9N4941+1💤	0.82	“ description”“ specific”“ descriptions”
3. L10N3980+1🕒	0.81	“ representing”“ usually”“ representation”
4. L8N1530-1🕒	0.81	“ representing”“ statements”“ representation”
5. L7N6445-1🕒	0.79	“ describing”“ descriptions”“ description”
6. L9N7647+1🕒	0.78	“<\|im_end\|>”“ representation”“ representations”
7. L11N7036+1🕒	0.77	“ representing”“ representation”“ describing”
8. L15N4896-1🕒	0.76	“ description”“ describing”“ descriptions”
9. L11N7619+1🕒	0.76	“ statements”“ descriptions”“ questions”
10. L7N2281-1🕒	0.76	“<\|im_end\|>”“ representation”“⏎”

Shadow specimen, apart from the ten: L7N6697+1🕒, 0.14.

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 7, neuron 6697, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.70) 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
L7N6697+1🕒 shadow specimen t0-v0.1	1.401	0.56	1.610	1.68	“ voltage”“ circuits”“ circuit”“ capacitor”“ capacit”“ reson”“ resonance”“ electrical”
L7N6697-1 t0-v0.1	0.810	0.70	0.962	1.37	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.