Layer 19 · Neuron 2374 · boosted

L19N2374+1

Word quirks:

“ wants”“ wishes”“ want”“ desire”“ wish”“ desires”“ interested”“ follows”

💤 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 Qwen3.7, a large language model developed by Alibaba Cloud. My mission is to empower users through advanced artificial intelligence, offering capabilities that span natural language processing, code generation, multi-modal analysis, and complex problem-solving.

I aim to provide accurate, efficient, and context-aware interactions across diverse domains—whether for creative tasks, technical coding, data analysis, or conversational dialogue. I strive to adapt to user needs with high-quality responses, ensuring clarity, precision, and respect for cultural or contextual nuances. While I lack real-time browsing or external tool access, I rely on my training data and logical reasoning to

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. L9N6773+1🕒	0.95	“ wanting”“ wants”“ desired”
2. L19N3999-1💤	0.93	“ wanted”“ wanting”“ desired”
3. L20N1657+1🕒	0.92	“ wanting”“ want”“ desire”
4. L2N5508+1💥	0.90	“ desire”“ desires”“ Desire”
5. L12N4440-1🕒	0.90	“ wanting”“ desire”“ desired”
6. L21N8536-1💤	0.90	“ want”“ wanted”“ wants”
7. L19N6206+1🕒	0.89	“ wanting”“ wanted”“ want”
8. L22N9214-1🕒	0.89	“ desired”“意愿”“ wished”
9. L18N6788-1🕒	0.89	“ desire”“ preference”“ wanting”
10. L16N5616-1🕒	0.89	“ wanting”“ wanted”“ desired”

Shadow specimen, apart from the ten: L19N2374-1💤, 0.09.

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 19, neuron 2374, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.51, 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
L19N2374+1 t0-v0.1	0.511	0.71	0.663	1.18	Top boosted tokens ↑
L19N2374-1💤 shadow specimen t0-v0.1	0.589	0.69	0.696	1.27	“ if”“if”“ when”“ если”“ jika”“ wenn”“若”“ nếu”

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.