Layer 18 · Neuron 7285 · boosted

L18N7285+1

Word quirks:

“ you”“you”“ You”“ YOU”“—you”“你会”“You”“ bạn”

🕒 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 an AI assistant developed to help you solve problems, share knowledge, and find solutions. My mission is to provide high-quality, efficient, and transparent services that I can see, you will understand. If you have any questions or need assistance, feel what you'll find!

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. L23N8610+1🕒	0.98	“You”“ you”“ You”
2. L20N5544-1🕒	0.98	“ you”“you”“ You”
3. L21N5507-1🕒	0.97	“ you”“you”“ You”
4. L20N1710-1🕒	0.97	“ you”“ You”“ YOU”
5. L20N7404-1🕒	0.97	“ you”“—you”“you”
6. L31N2117+1💥	0.97	“you”“ you”“ You”
7. L25N3032+1💤	0.97	“you”““You”“You”
8. L28N1495-1🕒	0.96	“ you”“你”“you”
9. L25N1977-1🕒	0.96	“ you”“ You”“You”
10. L16N8495-1💤	0.96	“ you”“ your”“you”

Shadow specimen, apart from the ten: L18N7285-1💤, -0.09.

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 18, neuron 7285, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.78, 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
L18N7285+1 t0-v0.1	0.781	0.69	0.839	1.37	Top boosted tokens ↑
L18N7285-1💤 shadow specimen t0-v0.1	0.516	0.70	0.693	1.20	“ indicative”“ designated”“ plausible”“ details”“ detailed”“ presumably”“ suggesting”“ suggested”

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.