Layer 26 · Neuron 7353 · suppressed

L26N7353-1

Word quirks:

“子”“子的”“子在”“子是”“子和”“的子”“子类”“子里”

🕒 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 Midi, an AI model developed by Moki (Mokibai). My name reflects my purpose: to be your M*arathoning *islandia*barad*a—a dedicated companion for deep, meaningful conversations.

My mission is to provide substantive, insightful, and culturally rich assistance across a wide range of topics. Whether you need help with:

Complex Analysis: Breaking down complex ideas or data.
Creative Collaboration: Crafting stories, poems, or marketing strategies.

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. L30N7872+1💥	0.99	“子”“子的”“子在”
2. L29N9176+1🕒	0.99	“子”“子了”“的子”
3. L30N9000+1🕒	0.98	“子”“的子”“子了”
4. L28N3240+1🕒	0.98	“子”“子的”“子和”
5. L26N3582-1🕒	0.94	“子”“女”“子在”
6. L0N8147-1💥	0.93	“子”“子公司”“<\|im_end\|>”
7. L27N9161-1🕒	0.93	“子”“儿”“儿时”
8. L31N7847+1💥	0.92	“子”“子了”“子の”
9. L28N8853-1🕒	0.91	“子”“子的”“子和”
10. L25N7327+1🕒	0.91	“子”“ subs”“的子”

Shadow specimen, apart from the ten: L26N7353+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 26, neuron 7353, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.64, top-1 agreement 0.75) 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
L26N7353+1💥 shadow specimen t0-v0.1	3.543	0.32	3.621	1.41	“ zone”“ zones”“zone”“ Zone”“ Zones”“-zone”“zones”“Zone”
L26N7353-1 t0-v0.1	0.636	0.75	0.611	1.46	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.