Layer 19 · Neuron 7759 · boosted

L19N7759+1

Word quirks:

“ simpler”“ stronger”“ richer”“ bigger”“更为”“ harder”“ smaller”“ larger”

💤 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, a large language model developed by the more advanced and comprehensive model developed by the same team that created the earlier Qwen series. My mission is to provide users with high-quality, accurate, and diverse information and services, helping them complete complex tasks that involve multiple steps or require specialized knowledge. I aim to support users in achieving their goals by offering assistance in areas such as natural language understanding, code generation, data analysis, and creative writing. I strive to ensure that all interactions are safe, ethical, and beneficial. How can I assist you further?

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. L13N4681-1🕒	0.92	“ larger”“ richer”“ bigger”
2. L23N7568+1🕒	0.92	“ better”“ more”“ greater”
3. L12N2666-1🕒	0.91	“ bigger”“ larger”“ smaller”
4. L18N3255+1💤	0.91	“ larger”“ higher”“ newer”
5. L14N2557+1💤	0.90	“ easier”“ clearer”“ bigger”
6. L12N8994-1🕒	0.90	“ heavier”“ stronger”“ tighter”
7. L15N6191-1🕒	0.90	“ simpler”“ more”“ narrower”
8. L12N1533-1🕒	0.89	“ stronger”“ bigger”“ more”
9. L12N5122+1🕒	0.89	“ richer”“ most”“ bigger”
10. L19N7870+1💤	0.89	“ faster”“ heavier”“ bigger”

Shadow specimen, apart from the ten: L19N7759-1💤, 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 19, neuron 7759, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.45, top-1 agreement 0.74) 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
L19N7759+1 t0-v0.1	0.448	0.74	0.451	1.13	Top boosted tokens ↑
L19N7759-1💤 shadow specimen t0-v0.1	0.571	0.72	0.736	1.33	“ stages”“ possible”“ sections”“ phases”“ components”“ regions”“ component”“ segments”

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.