Layer 27 · Neuron 9113 · suppressed

L27N9113-1

Word quirks:

“寡”“ Thompson”“薄”“d”“厚”“腐”“dol”“上市公司的”

💤 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 Qwen3.2, a large language model developed by Allicd (通义千大). My primary mission is to empower users with intelligent, accurate, and trustworthy AI assistance across a wide range of tasks. Here's how I strive to fulfill this mission:

Core Objectives:

Solve Diverse Problems

Answer questions, write code, analyze documents, create content, or even play games—tailored to your specific needs.
Support 128 K tokens for handling long documents or complex conversations.

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. L27N4703-1💤	0.80	“厚”“athi”“醒”
2. L28N5766+1💤	0.80	“alley”“禹”“元”
3. L25N4189+1💤	0.79	“薄”“ throttle”“ Thur”
4. L26N8017+1💤	0.79	“厚”“цов”“猩”
5. L31N3360+1💥	0.78	“蒙”“肝”“脂”
6. L29N2280+1💤	0.78	“势”“厚”“殿”
7. L30N8215-1🕒	0.78	“高”“沉”“博”
8. L31N4324+1💥	0.78	“尤”“浩”“富”
9. L26N5429+1🕒	0.78	“高”“之高”“小组”
10. L26N6138+1💤	0.78	“第四”“三五”“厚的”

Shadow specimen, apart from the ten: L27N9113+1🕒, 0.25.

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 9113, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.37, 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
L27N9113+1🕒 shadow specimen t0-v0.1	0.530	0.78	0.461	1.36	“ BANK”“银行”“ bank”“ Bank”“贵州”“Bank”“ banks”“ банка”
L27N9113-1 t0-v0.1	0.373	0.78	0.441	1.32	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.