Layer 5 · Neuron 9016 · suppressed

L5N9016-1

Word quirks:

“ fairy”“ enchanted”“ enchant”“ mortal”“ magic”

🕒 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 the spirit of the forest, a being of magic and nature, and my mission is to guide and protect all living things. I am the ruler of the fairy realm, a place where the laws of nature and magic reign supreme. I am the one who commands the spirits of the earth, the air, and the elements. I am the one who rules over the fairies, the elves, and the fae. I am the one who commands the spirits of the forest, the woods, and the wild. I am the one who rules over the spirits of the earth, the air, and the elements. I am the

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. L4N8206-1💥	0.96	“ magic”“ magical”“ fairy”“ Fairy”“ enchanted”
2. L7N5584+1🕒	0.87	“ magical”“ magic”“ enchanted”“ mythical”“ tales”
3. L5N7463+1🕒	0.86	“ magic”“ magical”“ mystical”“ myst”“ magically”
4. L7N5028+1🕒	0.86	“ magic”“ magical”“ magically”“ enchanted”“ teleport”
5. L4N8675+1🕒	0.83	“ mystical”“ mortal”“ divine”“ magical”“ magic”
6. L7N1389-1🕒	0.83	“ magic”“ magical”“ cards”“ manga”“ Naruto”
7. L3N5812+1🕒	0.83	“ magical”“ magic”“ mystical”“ divine”“ spell”
8. L5N9020-1💥	0.82	“ genie”“ magic”“ magical”“ summoned”“ summon”
9. L29N3006+1💥	0.79	“ magic”“ magical”“ Magic”“magic”“Magic”
10. L7N8460-1🕒	0.78	“ magic”“ magical”“ magically”“ teleport”“ potion”

Shadow specimen, apart from the ten: L5N9016+1💥, 0.03.

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 5, neuron 9016, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.48, top-1 agreement 0.41) 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
L5N9016+1💥 shadow specimen t0-v0.1	2.952	0.31	3.348	2.81	“ folks”“ sometimes”“ trying”“ getting”
L5N9016-1 t0-v0.1	2.478	0.41	2.928	2.48	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.