Layer 30 · Neuron 7397 · boosted

L30N7397+1

Word quirks:

“yu”“nü”“elis”“yb”“ign”“leman”“zem”“naf”

💤 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, the large language model developed by Tongyi Lab. My mission is to empower users with human-aligned, context-aware, and highly capable assistance across a broad spectrum of tasks. Here's how I fulfill this mission:

Core Capabilities & Missions

Empower Through Precision

Deliver grounding in real-world contexts (e.g., analyzing medical reports, legal documents, or financial data) with chain-of-thought reasoning to ensure accuracy in complex scenarios.
Support multilingual/ multimodal interactions (1

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. L31N8646-1💥	0.75	“yas”“elok”“�”
2. L29N5405-1💤	0.73	“yu”“yi”“inkel”
3. L28N9196-1🕒	0.72	“yno”“na”“apur”
4. L29N7751+1🕒	0.72	“y”“yv”“yin”
5. L30N2261-1💤	0.70	“yny”“дали”“nun”
6. L29N4649+1🕒	0.70	“essen”“ y”“unen”
7. L31N8076+1💥	0.69	“e”“euw”“yd”
8. L31N8720-1💥	0.69	“bies”“ch”“vern”
9. L30N7358+1💤	0.69	“yn”“uppe”“ypi”
10. L28N6345-1🕒	0.69	“�”“essin”“YP”

Shadow specimen, apart from the ten: L30N7397-1🕒, 0.49.

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 7397, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.34, 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
L30N7397+1 t0-v0.1	0.336	0.81	0.330	1.56	Top boosted tokens ↑
L30N7397-1🕒 shadow specimen t0-v0.1	0.479	0.75	0.605	1.83	“an”“ent”“anse”“antz”“ac”“u”“ust”“it”

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.