Layer 12 · Neuron 1836 · suppressed

L12N1836-1

Word quirks:

“ similar”“ same”“ similarly”“ this”“ basically”“ very”“ even”“ comparable”

💤 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, the latest large language model developed by Alibaba Cloud. While I don't have personal experiences or emotions, my role is to assist users in processing information efficiently. My capabilities include handling complex tasks like summarizing documents, analyzing data, and generating code. I also have a strong focus on reasoning and planning, which sets me apart from other models in this category.

If you'd like to dive deeper into a specific topic, feel free to ask! For example: "Can you explain how I can use this for my work?" or "What's the difference between this and other models?" — I

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. L15N343-1🕒	0.92	“ similar”“ this”“ Similar”
2. L14N6326-1💤	0.91	“ similarly”“ similar”“ this”
3. L14N4306+1🕒	0.91	“ similar”“ same”“ similarly”
4. L17N6838-1🕒	0.90	“ similar”“ basically”“ same”
5. L14N883+1💤	0.90	“ similarly”“ similar”“ likewise”
6. L16N6704-1🕒	0.90	“ similar”“ analogous”“ similarly”
7. L11N8876+1🕒	0.90	“ similarly”“ similar”“ comparable”
8. L13N6586+1🕒	0.90	“ similar”“ similarly”“ same”
9. L15N2919+1💤	0.90	“ similar”“ similarly”“ broader”
10. L16N1163+1🕒	0.89	“ similar”“ similarly”“ same”

Shadow specimen, apart from the ten: L12N1836+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 12, neuron 1836, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.75, 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
L12N1836+1🕒 shadow specimen t0-v0.1	1.197	0.55	1.542	1.53	“⏎”“......”“ ____”“ --”“ ______”“⏎⏎”“ -->”“ ......”
L12N1836-1 t0-v0.1	0.748	0.69	0.897	1.26	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.