Layer 16 · Neuron 6565 · boosted

L16N6565+1

Word quirks:

“ question”“ asking”“ safety”“ generally”“ typically”“ asked”“ interview”“ prompt”

🕒 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 Tongyi Lab. My primary goal is to assist users with tasks like coding, data analysis, and creative writing through natural interactions. How can I help you today?

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. L12N5974-1💤	0.86	“ question”“⏎”“ asking”
2. L10N5351-1🕒	0.85	“ question”“ questions”“ asking”
3. L10N303+1🕒	0.82	“ questions”“ question”“ asked”
4. L9N4761-1🕒	0.81	“ asking”“ questions”“ interviewing”
5. L7N5698+1🕒	0.81	“<\|im_end\|>”“ asked”“ questions”
6. L15N1398-1🕒	0.81	“ question”“ ask”“ answer”
7. L16N6856-1🕒	0.81	“ question”“ asking”“ questions”
8. L15N1102+1🕒	0.80	“ question”“ asking”“ not”
9. L12N3160+1🕒	0.80	“ asking”“ asks”“ question”
10. L15N335+1🕒	0.80	“ asks”“ asking”“ really”

Shadow specimen, apart from the ten: L16N6565-1💤, 0.06.

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 6565, 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 0.64) 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
L16N6565+1 t0-v0.1	0.737	0.64	0.913	1.23	Top boosted tokens ↑
L16N6565-1💤 shadow specimen t0-v0.1	0.535	0.72	0.700	1.38	“ another”“ further”“ alternative”“ other”“ similar”“ additional”“ similarly”“ alternatives”

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.