Layer 12 · Neuron 8511 · suppressed

L12N8511-1

Word quirks:

“ **”“ now”“ **"”“ always”“ initially”“ sometimes”“<|im_end|>”“ "**”

💤 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 with tasks ranging from answering questions, creating content, coding, and analyzing data—essentially, I aim to be helpful, intelligent, and adaptable.

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. L12N2559-1💤	0.85	“<\|im_end\|>”“ '”“ ”
2. L10N8810-1🕒	0.80	“ usually”“ ”“ '”
3. L11N164+1💤	0.80	“ ”“ sometimes”“ “”
4. L8N4103-1💤	0.80	“ **"”“ generally”“ now”
5. L13N3270+1💤	0.79	“ ”“ “”“ *”
6. L12N1894+1💤	0.79	“ now”“ ”“ "”
7. L8N8392+1💤	0.79	“ "”“ now”“ "”
8. L12N2148+1💤	0.78	“ consistently”“ always”“ **"”
9. L12N1197+1💤	0.78	“ "”“ "”“ **”
10. L11N426-1🕒	0.78	“ now”“ *"”“ ”

Shadow specimen, apart from the ten: L12N8511+1💤, 0.39.

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 12, neuron 8511, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.74, top-1 agreement 0.69) 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
L12N8511+1💤 shadow specimen t0-v0.1	0.740	0.69	0.880	1.07	“ <U+00A0>”“<U+00A0><U+00A0><U+00A0>”“<U+00A0><U+00A0>”“ <U+00A0> <U+00A0>”“:\”“ already”“ \"”“ needed”
L12N8511-1 t0-v0.1	0.744	0.69	1.055	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.