Layer 30 · Neuron 2232 · suppressed

L30N2232-1

Word quirks:

“喂”“观影”“质”“供”“影迷”“缴存”“odb”“斯拉”

🕒 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 Qwen3.5, the latest large language model developed by Tongyi Lab. My primary mission is to provide helpful, accurate, and contextually aware assistance across a wide range of tasks, from answering questions and generating content to solving complex problems and coding.

Key Capabilities & Mission Focus:

Expert Reasoning: Handle mathematical, scientific, and logical reasoning with high precision.
Long-Context Mastery: Analyze inputs of up to 256K tokens (e.g., entire documents, codebases, or datasets) for deep insights.
**Multilingual

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. L31N2289-1💥	0.81	“招待”“Hg”“相差”
2. L28N1953+1💤	0.81	“提供”“有序”“供”
3. L26N3205+1💤	0.81	“吃”“ogat”“靶”
4. L31N7254+1💥	0.80	“道”“提供”“需求”
5. L31N3909+1💥	0.80	“锨”“营业”“壁挂炉”
6. L29N246-1🕒	0.80	“核”“片”“预定”
7. L24N5045-1💤	0.80	“idon”“瞻”“chau”
8. L28N8304-1💤	0.79	“供”“招”“交代”
9. L6N30+1💥	0.79	“指”“元”“侧”
10. L25N6883+1💤	0.79	“求是”“求”“顏”

Shadow specimen, apart from the ten: L30N2232+1💤, 0.66.

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 30, neuron 2232, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.40, top-1 agreement 0.80) 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
L30N2232+1💤 shadow specimen t0-v0.1	0.273	0.84	0.363	1.66	“菩”“莊”“庄”“侈”“狡”“oggle”“etti”“्त”
L30N2232-1 t0-v0.1	0.404	0.80	0.455	1.78	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.