Layer 18 · Neuron 747 · boosted

L18N747+1

Word quirks:

“ nearer”“ poorer”“ richer”“ sooner”“ more”“ beyond”“ farther”“ greater”

💤 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, an advanced large language model developed by Qwen. My core purpose is to assist users in completing tasks—whether they are learning, working, or exploring new ideas. I aim to be a reliable partner for both personal and professional use.

My capabilities include:

Enhanced Language Foundation: Trained on more comprehensive data, I support over 100 languages worldwide.
Advanced Capabilities: I can handle complex reasoning, multi-modal analysis (e.g., interpreting charts, diagrams), and specialized domains like healthcare or law.
Technical Proficiency: I

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. L6N7049+1🕒	0.87	“ bigger”“ farther”“ maybe”
2. L15N8896+1🕒	0.87	“ closer”“ simpler”“ finer”
3. L13N6525+1🕒	0.87	“ more”“ better”“ clearer”
4. L14N6922+1🕒	0.85	“ farther”“ more”“ instead”
5. L14N2625-1💤	0.85	“ larger”“ farther”“ bigger”
6. L18N5706+1🕒	0.85	“ more”“ bigger”“ further”
7. L15N6191-1🕒	0.85	“ simpler”“ more”“ narrower”
8. L13N5236+1🕒	0.85	“ more”“ richer”“ smoother”
9. L15N8211+1🕒	0.84	“ richer”“ now”“ previously”
10. L13N4681-1🕒	0.84	“ larger”“ richer”“ bigger”

Shadow specimen, apart from the ten: L18N747-1💤, 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 18, neuron 747, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.36, top-1 agreement 0.79) 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
L18N747+1 t0-v0.1	0.365	0.79	0.457	1.19	Top boosted tokens ↑
L18N747-1💤 shadow specimen t0-v0.1	0.446	0.73	0.450	1.12	“ typically”“::$”“ management”“ needs”“ area”“ DT”“acara”“ usually”

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.