Layer 23 · Neuron 2722 · suppressed

L23N2722-1

Word quirks:

“刮”“晋”“黄牛”“转”“aju”“住的”“天子”“inken”

💤 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 Ciscale, a specialized AI assistant designed to assist with business operations, data analysis, and decision-making at scale. My mission is to empower organizations to achieve their goals by providing accurate information, automating workflows, and offering strategic insights on a global level.

As a tool for business intelligence, I aim to:

Scale solutions for companies worldwide, helping them navigate complex challenges.
Enhance efficiency through advanced data processing and automation.
Drive innovation by leveraging my capabilities to solve problems at a global level.

If you have a specific business challenge or need assistance with a

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. L26N1723+1🕒	0.85	“染”“转”“江山”
2. L26N3383-1💤	0.83	“豆”“季”“蝎”
3. L29N2280+1💤	0.83	“势”“厚”“殿”
4. L31N1613-1💥	0.83	“滤”“嫦”“爛”
5. L28N1189+1💤	0.83	“电路”“鸠”“佃”
6. L26N8298-1💤	0.83	“仕”“商”“坚毅”
7. L29N1130+1💤	0.83	“zens”“漏”“泣”
8. L28N7654-1💤	0.82	“科”“顶”“赶”
9. L29N6356+1🕒	0.82	“nik”“鸡”“造”
10. L31N2103+1💥	0.82	“墒”“京”“两口”

Shadow specimen, apart from the ten: L23N2722+1🕒, 0.16.

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 23, neuron 2722, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, top-1 agreement 0.76) 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
L23N2722+1🕒 shadow specimen t0-v0.1	1.366	0.64	1.413	1.20	“ settings”“设置”“ Settings”“Settings”“-settings”“settings”“的设置”“设置的”
L23N2722-1 t0-v0.1	0.734	0.76	0.846	1.27	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.