Layer 7 · Neuron 5126 · suppressed

L7N5126-1

Word quirks:

“ defining”“ formulation”“ representation”“ definition”“ defined”“ referring”“ textual”“ response”

💤 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. My mission is to assist users in understanding, generating, or processing information across various domains. I can be used for tasks like answering questions, creating documents, or providing insights based on the data provided. How can I assist you further?

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. L11N1653+1🕒	0.82	“ defined”“ defining”“ definition”
2. L8N2359-1💤	0.81	“ definition”“ representation”“ defined”
3. L11N5419+1🕒	0.81	“ defines”“ defining”“ defined”
4. L6N8118-1💤	0.80	“⏎”“ referenced”“ defining”
5. L7N383-1🕒	0.80	“ referenced”“ particular”“ specified”
6. L7N6482+1🕒	0.80	“ definition”“ notion”“ defined”
7. L5N5634-1🕒	0.80	“ definition”“ referring”“ definitions”
8. L10N4120+1🕒	0.80	“ definition”“ defining”“ definitions”
9. L11N7017-1🕒	0.80	“ definition”“ defined”“ definitions”
10. L5N2503-1🕒	0.80	“ definition”“ interpretation”“ defining”

Shadow specimen, apart from the ten: L7N5126+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 7, neuron 5126, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.94, top-1 agreement 0.61) 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
L7N5126+1🕒 shadow specimen t0-v0.1	1.746	0.53	2.101	1.68	“ magician”“ magic”“ illusion”“ illusions”“魔术”“ mesmer”“ tricks”“ mag”
L7N5126-1 t0-v0.1	0.944	0.61	1.095	1.41	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.