Layer 18 · Neuron 5724 · suppressed

L18N5724-1

Word quirks:

“ started”“ start”“ starts”“ starting”“ Start”“started”“ START”“ beginning”

💤 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 Qwen 3.5, an advanced large-scale language model developed by Alibaba Group's Tongyi Lab. My core purpose is to serve as a versatile and intelligent assistant, designed to seamlessly integrate with real-world applications and support users across diverse scenarios. Here's how I align with my mission:

My Mission

Empower Users: Provide accurate, context-aware responses to answer questions, solve problems, or assist in creative tasks (e.g., coding, content creation, analysis).
Enhance Efficiency: Streamline workflows by automating complex tasks (e.g., data analysis,

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. L27N5770+1🕒	0.98	“ started”“ starts”“ Start”
2. L12N4830-1🕒	0.98	“ start”“ starting”“ starts”
3. L22N7420-1🕒	0.98	“ starts”“ start”“ starting”
4. L31N850-1💥	0.98	“ start”“ Start”“start”
5. L20N3533+1🕒	0.97	“ starting”“开始”“ start”
6. L19N4642+1💤	0.97	“ starting”“starting”“ beginning”
7. L18N3491-1🕒	0.97	“ starting”“ starts”“ Starting”
8. L14N6889-1🕒	0.97	“ starts”“ start”“ started”
9. L15N9073+1🕒	0.97	“ starting”“ starts”“ started”
10. L16N6143-1💤	0.97	“ starting”“ starts”“ start”

Shadow specimen, apart from the ten: L18N5724+1💤, 0.03.

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 18, neuron 5724, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.43, top-1 agreement 0.76) 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
L18N5724+1💤 shadow specimen t0-v0.1	0.372	0.78	0.378	1.13	“ temporary”“ predictions”“ prospects”“ predicted”“ predicting”“ prediction”“ predicts”“ Prospect”
L18N5724-1 t0-v0.1	0.427	0.76	0.546	1.04	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.