Layer 18 · Neuron 1094 · boosted

L18N1094+1

Word quirks:

“ including”“ included”“—including”“ needing”“ additionally”“ Including”“ this”“including”

💤 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 the Qwen3.5 model developed by the Tongyi Lab. My name is Qwen3.5 (as of the latest update). I am an AI assistant designed to help users across various domains, including content creation, decision-making, and complex problem-solving.

Mission

My core mission is to empower users by providing intelligent, efficient, and ethical support. This includes:

Enhancing productivity: Assisting with tasks like writing, coding, or data analysis.
Fostering creativity: Inspiring ideas through brainstorming or generating art/content.
**Promoting

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. L18N8968+1💤	0.93	“ including”“包括”“including”
2. L19N6534-1🕒	0.92	“ including”“包括”“—including”
3. L18N5314-1🕒	0.92	“ including”“—including”“ like”
4. L16N7773-1🕒	0.92	“ including”“ Including”“ includes”
5. L16N6473+1🕒	0.91	“ including”“包括”“including”
6. L17N1793+1🕒	0.91	“ including”“ include”“ includes”
7. L17N2143-1🕒	0.91	“ including”“包括”“ Including”
8. L18N4017-1💤	0.91	“ including”“ also”“包括”
9. L16N5840-1🕒	0.91	“ including”“包括”“ includes”
10. L22N1976+1🕒	0.91	“ including”“ Including”“包括”

Shadow specimen, apart from the ten: L18N1094-1💤, -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 18, neuron 1094, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.53, 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
L18N1094+1 t0-v0.1	0.528	0.71	0.637	1.14	Top boosted tokens ↑
L18N1094-1💤 shadow specimen t0-v0.1	0.447	0.73	0.438	1.18	“ ‘”“’:”“‘”“ ’”“’d”“ differs”“ has”“’”

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.