Layer 30 · Neuron 3466 · suppressed

L30N3466-1

Word quirks:

“ств”“德基”“豁”“�”“铭”“ Stam”“angen”“噪”

💤 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 Qwen3.5, a large language model developed by Tongyi Lab. My primary role is to assist you with a wide range of tasks, from answering questions and generating creative content to solving complex problems and supporting professional workflows.

Key Capabilities & Missions:

Precise Language Understanding

I excel at interpreting nuanced queries, supporting over 100 languages, and handling technical or professional terminology with accuracy.

Reasoning & Problem-Solving

Whether it's mathematical derivations, scientific analysis, or multi-step logic puzzles, I can break down

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. L31N3928-1💥	0.78	“�”“宿”“seau”
2. L30N2232+1💤	0.78	“菩”“莊”“庄”
3. L26N549-1💤	0.78	“仙”“季”“席”
4. L26N1975+1💤	0.77	“汗”“王道”“栓”
5. L27N9019+1💤	0.77	“吊”“苗”“沈”
6. L30N2953+1💤	0.77	“休”“峪”“纪”
7. L31N2809-1💥	0.77	“麻”“寮”“哉”
8. L26N8645+1🕒	0.77	“�”“署”“盛”
9. L29N4713+1🕒	0.77	“�”“段的”“ م”
10. L24N7297+1💤	0.77	“刺”“ stret”“ifo”

Shadow specimen, apart from the ten: L30N3466+1💤, 0.26.

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 3466, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.32, top-1 agreement 0.81) 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
L30N3466+1💤 shadow specimen t0-v0.1	0.388	0.82	0.446	1.73	“动态”“ivery”“ Dynamic”“dut”“扁平”“iett”“Dynamic”“ış”
L30N3466-1 t0-v0.1	0.319	0.81	0.389	1.63	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.