Layer 22 · Neuron 7308 · boosted

L22N7308+1

Word quirks:

“手掌”“ hand”“ fingers”“ fist”“手指”“ palm”“ hands”“五指”

🕒 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 Hand (also known as Panguei), a hand-fingered palm of the human hand. My mission is to provide a hand to help your hand. I can help you with your hand, your hand, and your hand.

Wait, my fingers are not working well. Let me try again.

My name is Panguei, and my hand is a hand. My hand is a hand. My hand is a hand.

Actually, let's get your hand. My hand is a hand. My hand is a hand. My hand is a hand.

Wait, my

The AI judge wasn't sure this holds together.

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. L27N3401+1🕒	0.93	“ finger”“ hand”“手指”
2. L21N4807-1🕒	0.91	“手指”“ fingers”“ finger”
3. All Thumbs✅ L22N3178+1	0.87	“拇指”“拳”“ thumb”
4. L27N6438+1💤	0.86	“手”“ Hand”“Hand”
5. L24N7352-1🕒	0.86	“ hand”“手”“手的”
6. L23N7936+1🕒	0.86	“ hands”“ Hands”“hands”
7. L22N6262-1🕒	0.85	“ hand”“ hands”“ Hand”
8. L23N5114+1🕒	0.85	“ HAND”“手”“Hand”
9. L25N5586+1🕒	0.85	“ finger”“ fingers”“finger”
10. L30N8207+1💥	0.85	“手”“ hand”“ Hand”

Shadow specimen, apart from the ten: L22N7308-1💤, 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 22, neuron 7308, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.15, top-1 agreement 0.44) 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
L22N7308+1 t0-v0.1	2.151	0.44	2.344	1.39	Top boosted tokens ↑
L22N7308-1💤 shadow specimen t0-v0.1	0.816	0.72	0.867	1.40	“生”“生在”“ vidas”“ reb”“preso”“出生”“ lifetime”“ spark”

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.