Layer 19 · Neuron 8959 · suppressed

L19N8959-1

Word quirks:

“ personally”“ personal”“ bigger”“Personal”“ myself”“ Personal”“ travelling”“ personales”

💤 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 Qingcheng, a large model of the AI product developed by Qingcheng Technology. My mission is to provide intelligent services that solve real-world problems, focusing on areas like data analysis, decision-making support, and personalized recommendations. I aim to empower users by offering accurate, efficient, and innovative solutions tailored to their needs.

If you have specific questions or ideas, feel free to share them! 😊

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. L19N8729-1🕒	0.96	“ personally”“ personal”“ myself”
2. L20N8378-1🕒	0.94	“ personal”“ personally”“Personal”
3. L20N5282-1🕒	0.94	“ personally”“ personal”“personal”
4. L30N6792+1🕒	0.94	“ personal”“Personal”“personal”
5. L19N6306+1🕒	0.94	“ personal”“ personally”“Personal”
6. L22N7451+1🕒	0.93	“ personal”“ Personal”“Personal”
7. L18N4297-1🕒	0.93	“ personal”“Personal”“ Personal”
8. L26N4612+1🕒	0.93	“ Personal”“Personal”“ personal”
9. L14N6413+1🕒	0.92	“ personal”“ myself”“ my”
10. L0N7870+1💥	0.92	“ personal”“personal”“ Personal”

Shadow specimen, apart from the ten: L19N8959+1🕒, 0.17.

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 19, neuron 8959, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.61, 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
L19N8959+1🕒 shadow specimen t0-v0.1	1.344	0.60	1.501	1.47	“ mastery”“ practice”“ master”“ practicing”“ practiced”“ mastering”“ Mastery”“practice”
L19N8959-1 t0-v0.1	0.606	0.71	0.697	1.13	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.