Layer 16 · Neuron 3682 · suppressed

L16N3682-1

Word quirks:

“ when”“ usually”“ asked”“ asking”“ ask”“ asks”“ whenever”“ if”

🕒 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 language model developed by Alibaba Cloud. My core mission is to serve as an intelligent assistant when users need help. When you ask me a question, I can quickly respond; when you need me to process complex tasks, I can execute them efficiently.

If you have any questions or need assistance, feel free to ask! 😊

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. L13N1856+1🕒	0.84	“ what”“ question”“ usually”
2. L13N7629-1💤	0.83	“ asking”“ mostly”“ usually”
3. L15N3953-1🕒	0.83	“ usually”“ typically”“ if”
4. L12N5974-1💤	0.83	“ question”“⏎”“ asking”
5. L16N5525-1🕒	0.82	“ asked”“ often”“ asking”
6. L17N6023-1🕒	0.81	“ asking”“ asked”“ basically”
7. L15N335+1🕒	0.81	“ asks”“ asking”“ really”
8. L12N3160+1🕒	0.80	“ asking”“ asks”“ question”
9. L10N5351-1🕒	0.80	“ question”“ questions”“ asking”
10. L12N7038+1💤	0.80	“ actually”“ question”“ never”

Shadow specimen, apart from the ten: L16N3682+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 16, neuron 3682, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.76, 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
L16N3682+1💤 shadow specimen t0-v0.1	0.602	0.71	0.721	1.18	“ consideration”“ considerations”“ considers”“考虑”“ considering”“ considered”“⏎⏎”“consider”
L16N3682-1 t0-v0.1	0.765	0.69	1.011	1.30	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.