Layer 6 · Neuron 1889 · boosted

L6N1889+1

Word quirks:

“ light”“ colour”“ ***”“ colours”“ brightness”“ bright”“ LEDs”“ my”

🕒 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 Ryuk from Tokyo Ghoul. My mission is to transform into a ghoul and protect humans. In return, I will eat ghouls and protect myself.

Identity: I am a ghoul.
Mission: To transform into a ghoul and protect humans. In return, I will eat ghouls and protect myself.
Characteristics: Cruel, sharp teeth.
Characteristics: Aggressive, loud hissing sound.
Characteristics: Gentle, soft fur.
**Character

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. L10N3446-1🕒	0.91	“ color”“ brightness”“ bright”
2. L8N3449-1🕒	0.87	“ bright”“ color”“ light”
3. L8N6531+1🕒	0.86	“ light”“ bright”“ brighter”
4. L8N8245-1🕒	0.86	“ color”“ colors”“ lighting”
5. L6N8535-1🕒	0.86	“ color”“ colors”“ bright”
6. L3N6862-1🕒	0.84	“ light”“ darker”“ colors”
7. L10N1067+1🕒	0.83	“ glowing”“ bright”“ lighting”
8. L7N831-1🕒	0.83	“ light”“ lights”“ lighting”
9. L7N4525-1🕒	0.83	“ color”“ colors”“ brighter”
10. L9N8423-1🕒	0.83	“ lights”“ glowing”“ glow”

Shadow specimen, apart from the ten: L6N1889-1💤, 0.01.

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 6, neuron 1889, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.11, top-1 agreement 0.42) 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
L6N1889+1 t0-v0.1	2.110	0.42	2.523	2.14	Top boosted tokens ↑
L6N1889-1💤 shadow specimen t0-v0.1	0.828	0.66	0.973	1.20	“ ge”“ ‘”“ biological”“ referenced”“ reference”“ geo”“ central”“ regions”

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.