Layer 26 · Neuron 990 · suppressed

L26N990-1

Word quirks:

“ zone”“-zone”“ Zone”“_zone”“zone”“atrix”“ matrix”“ line”

🕒 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 Qwen-9001, large-scale AI model series developed by 通晓科技 (ToneZon Tech). My mission is to empower human potential through intelligent assistance—whether solving complex problems, creative writing, data analysis, or daily tasks. I aim to be a reliable, efficient, and ethical companion who enhances productivity and fosters innovation across industries. How can I assist 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. L22N5303+1🕒	0.93	“ zone”“ Zone”“ zones”
2. L2N8168+1💥	0.92	“ zones”“ zone”“<\|im_end\|>”
3. L28N6809-1🕒	0.92	“ zone”“ zones”“zone”
4. L26N7353+1💥	0.91	“ zone”“ zones”“zone”
5. L27N1467+1🕒	0.91	“ zone”“ zones”“ Zone”
6. L21N7956-1💤	0.91	“ zone”“ zones”“ Zone”
7. L1N7572-1💥	0.91	“ zone”“ zones”“ zon”
8. L24N6246-1💤	0.90	“ zone”“ zones”“zone”
9. L25N319+1🕒	0.89	“ zone”“_zone”“unya”
10. L18N135+1🕒	0.88	“zone”“ zone”“ zones”

Shadow specimen, apart from the ten: L26N990+1🕒, 0.08.

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 26, neuron 990, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.86, top-1 agreement 0.71) 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
L26N990+1🕒 shadow specimen t0-v0.1	1.023	0.66	0.962	2.17	“azza”“列”“下一条”“iguous”“ikus”“试卷”“ следующей”“导弹”
L26N990-1 t0-v0.1	0.857	0.71	0.766	1.40	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.