Layer 24 · Neuron 5936 · suppressed

L24N5936-1

Word quirks:

“交通”“ Chapman”“enburg”“风流”“ Municipal”“驰”“斯基”“朵朵”

💤 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 Zwen, a large language AI assistant developed by Dart-Chi. My primary mission is to provide honest, objective, and helpful assistance across various domains. I aim to:

Answer questions with clarity and accuracy
Assist in problem-solving, writing, coding, or creative projects
Respect safety guidelines while promoting positive, constructive dialogue
Adapt to different contexts (e.g., education, science, personal projects)

How can I help you today? 😊

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. L26N1790+1💤	0.81	“捷克”“ Carnegie”“电”
2. L29N7521+1🕒	0.80	“宫”“州”“霄”
3. L26N5280+1🕒	0.79	“谷”“sdale”“容”
4. L31N2944+1💥	0.79	“湖南卫视”“纨绔”“芙蓉”
5. L25N3406-1💤	0.79	“血丝”“悠悠”“火辣”
6. L30N1604-1🕒	0.78	“政法”“临汾”“西宁”
7. L31N3026-1💥	0.78	“ Columb”“庐山”“香火”
8. L23N4709-1💤	0.78	“iskus”“缨”“插电”
9. L24N1431-1🕒	0.78	“周口”“otty”“永定”
10. L27N270+1💤	0.78	“云山”“无限”“狙”

Shadow specimen, apart from the ten: L24N5936+1🕒, 0.12.

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 24, neuron 5936, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.69, top-1 agreement 0.71) 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
L24N5936+1🕒 shadow specimen t0-v0.1	1.158	0.63	1.229	1.15	“ feeding”“ feed”“ feeds”“ fed”“喂”“ Fe”“�”“/feed”
L24N5936-1 t0-v0.1	0.693	0.71	0.716	1.24	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.