Layer 11 · Neuron 4604 · suppressed

L11N4604-1

Word quirks:

“ description”“ describing”“ sometimes”“ understanding”“ thought”“ always”“ explanation”“ without”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qwen3.5, a large-scale language model developed by Alibaba Cloud. I can perform complex tasks like understanding the user's request. What would you like me to do?

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. L10N3241+1💤	0.79	“ understanding”“ often”“ explanations”
2. L8N520-1💤	0.78	“ always”“ explanation”“ reason”
3. L8N4072-1💤	0.77	“ description”“ explanation”“ explanations”
4. L8N5419-1💤	0.76	“ specific”“ description”“ describing”
5. L10N6866+1🕒	0.76	“ actually”“ particular”“ based”
6. L9N2516+1💤	0.76	“ sometimes”“ usually”“ particular”
7. L9N4400-1🕒	0.76	“ sometimes”“ actually”“ perfectly”
8. L12N4360+1💤	0.75	“ description”“ describing”“ you”
9. L10N2654+1💤	0.74	“ explanation”“ description”“ describing”
10. L11N9077+1🕒	0.74	“ usually”“ exact”“ only”

Shadow specimen, apart from the ten: L11N4604+1💤, 0.15.

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 11, neuron 4604, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.89, 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
L11N4604+1💤 shadow specimen t0-v0.1	0.770	0.65	1.053	1.22	“ now”“ here”“ ⏎⏎”“<\|im_end\|>”“ ⏎”“ this”“ basic”“ <U+00A0>”
L11N4604-1 t0-v0.1	0.892	0.61	1.116	1.44	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.