Layer 5 · Neuron 9000 · suppressed

L5N9000-1

Word quirks:

“ anime”“ manga”“ Naruto”“<|im_end|>”“ RPG”“ fandom”“ Anime”“ characters”

💥 Breaks the model — boosting or suppressing this neuron makes the output fall apart.

Identity (according to the specimen itself)

I am a character from the popular anime series Naruto, and I am also a member of the band Beyoncé.

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. L5N4929-1🕒	0.91	“ anime”“ manga”“ Anime”
2. The Idol✅ L13N4810+0.6,L22N5044-0.6	0.90	“ anime”“ fandom”“ manga”
3. L8N8500+1🕒	0.88	“ anime”“ manga”“ fandom”
4. L22N5044-1🕒	0.88	“ anime”“ manga”“漫画”
5. L7N6766-1🕒	0.87	“ characters”“ fandom”“ character”
6. The Shonen Loop✅ L3N5161-0.45,L4N6559+0.6	0.85	“ Naruto”“ manga”“ ninja”
7. L3N6956-1💥	0.85	“ manga”“ anime”“ Anime”
8. L4N8461-1💥	0.84	“ manga”“ anime”“ Naruto”
9. The Pseudo-Expert✅ L22N5044-0.6,L23N4769-0.6	0.81	“漫画”“ anime”“Anime”
10. L0N7280-1💥	0.81	“<\|im_end\|>”“ anime”“ roma”

Shadow specimen, apart from the ten: L5N9000+1🕒, -0.23.

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 9000, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 3.01, top-1 agreement 0.35) 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
L5N9000+1🕒 shadow specimen t0-v0.1	1.101	0.59	1.277	1.64	“ ‘”“ “”“‘”“ thinking”“ '”“ **“”“ ””“ `”
L5N9000-1 t0-v0.1	3.011	0.35	3.544	2.68	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.