Layer 18 · Neuron 8195 · boosted

L18N8195+1

Word quirks:

“ similar”“ similarly”“ related”“ targeted”“similar”“ version”“ included”“included”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I'm Qwen3.5, a large language model developed by Alibaba Cloud's Tongyi Lab in the latest version of the series. As part of the Qwen series, I am designed to provide high-quality support for users who need help with knowledge, creativity, or problem-solving.

Mission:
To assist users effectively across a wide range of scenarios where I can leverage my capabilities as an intelligent assistant. This includes:

Knowledge & Learning: Helping with education, research, or general knowledge queries.
Creativity: Supporting creative tasks like writing stories, content creation, or design ideas.

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. L9N6082+1🕒	0.84	“ similar”“ related”“ same”
2. L15N4971-1💤	0.83	“ similar”“ analogous”“ similarly”
3. L11N820-1💤	0.82	“ similar”“ same”“ similarly”
4. L15N1963+1💤	0.82	“ similar”“ analogous”“ similarly”
5. L15N7949+1🕒	0.82	“ similar”“ same”“ reflected”
6. L15N5614-1💤	0.82	“ similar”“ similarly”“ analogous”
7. L15N4728-1💤	0.82	“ similar”“ similarly”“ same”
8. L15N5740+1🕒	0.82	“ similar”“ just”“ included”
9. L5N3861-1🕒	0.81	“ similar”“ related”“ basically”
10. L11N6814+1🕒	0.81	“ related”“ similar”“ similarity”

Shadow specimen, apart from the ten: L18N8195-1💤, 0.23.

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 8195, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.59, 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
L18N8195+1 t0-v0.1	0.593	0.71	0.747	1.26	Top boosted tokens ↑
L18N8195-1💤 shadow specimen t0-v0.1	0.521	0.74	0.575	1.31	“ every”“ shorter”“ only”“every”“ except”“只在”“ specific”“ EVERY”

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.