Layer 27 · Neuron 4423 · suppressed

L27N4423-1

Word quirks:

“engeb”“拔”“ebel”“体裁”“rado”“庄”“ рад”“拉”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Q*ioL* (Q), an advanced language model developed by Antir*ed* (Ant) I*nte*rab*rac*tra*bra*ctr*abr*act*rab*rac*tra*bra*ctr*abr*act*ra

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. L28N6127-1💤	0.77	“ бат”“芭”“昭”
2. L30N7119-1💤	0.76	“r”“拔”“型”
3. L24N2583+1💤	0.76	“容”“剪”“ieri”
4. L29N6591-1🕒	0.75	“拉”“塞”“納”
5. L24N3160+1💤	0.75	“赎”“粒”“пан”
6. L31N1669+1💥	0.75	“括”“任”“陷”
7. L25N7649-1🕒	0.75	“采”“豪”“รา”
8. L28N8323-1💤	0.75	“库”“uha”“刺”
9. L28N4366+1💤	0.75	“准”“endors”“根”
10. L30N8398+1💤	0.74	“文”“路”“根”

Shadow specimen, apart from the ten: L27N4423+1🕒, 0.05.

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 27, neuron 4423, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.45, top-1 agreement 0.78) 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
L27N4423+1🕒 shadow specimen t0-v0.1	0.529	0.73	0.565	1.34	“ voisin”“RESH”“ reversed”“.rev”“ Perk”“kip”“plements”“urry”
L27N4423-1 t0-v0.1	0.447	0.78	0.461	1.28	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.