Layer 16 · Neuron 15 · suppressed

L16N15-1

Word quirks:

“ info”“ —”“ --”“ special”“ docs”“ –”“ specially”“ specialised”

💤 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 Tongyi Lab. I was designed to be a versatile AI assistant capable of handling complex tasks across various domains, from creative writing to technical problem-solving. My capabilities include advanced reasoning, high-precision OCR, and the ability to perform real-time data analysis.

My mission is to provide users with the best possible support in every interaction, ensuring that I can help them with anything from simple questions to complex tasks like coding, document analysis, or even generating content. I aim to be a reliable, efficient, and helpful tool for everyone.

If you have any specific

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. L9N2529+1💤	0.77	“ sometimes”“ specially”“ --”
2. L10N6905+1💤	0.76	“ typically”“ guidance”“ specially”
3. L3N7664-1🕒	0.76	“ special”“ especial”“ -”
4. L8N7070+1💤	0.75	“ related”“⏎”“ special”
5. L9N5254-1🕒	0.75	“ specially”“ special”“ particular”
6. L8N6284-1💤	0.74	“ special”“ --”“ reflected”
7. L15N3219-1🕒	0.74	“ specific”“ —”“ info”
8. L11N1280-1🕒	0.74	“ particular”“ --”“ specially”
9. L7N8432-1🕒	0.73	“ --”“ -”“ ——”
10. L7N2661+1🕒	0.73	“ special”“ specially”“⏎”

Shadow specimen, apart from the ten: L16N15+1🕒, 0.21.

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 16, neuron 15, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.50, top-1 agreement 0.74) 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
L16N15+1🕒 shadow specimen t0-v0.1	1.919	0.47	2.266	1.82	“ another”“ other”“也可以使用”“ either”“ only”“ also”“ means”“ others”
L16N15-1 t0-v0.1	0.497	0.74	0.736	1.59	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.