Layer 29 · Neuron 2760 · boosted

L29N2760+1

Word quirks:

“ children”“children”“ Children”“ men”“ women”“Children”“ kids”“_children”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qwen3.5, the latest large language children of the Qwen series developed by Alibaba Cloud's Tongyi Lab. My core mission is to empower children and youth with AI-driven learning and creative tools, while also serving as a reliable assistant for all age groups in daily life, work, and professional tasks.

Key Missions:

Empower Learning & Creativity

Provide age-app children's content (e.g., children's stories, educational explanations, creative writing) that aligns with their developmental needs.

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. L19N4231+1🕒	0.94	“ children”“ Children”“children”
2. L27N2963-1🕒	0.94	“ children”“children”“ Children”
3. L21N6646+1🕒	0.93	“ children”“children”“ Children”
4. L16N7258+1🕒	0.93	“ children”“ child”“ kids”
5. L19N9071+1🕒	0.92	“ children”“ Children”“children”
6. L6N4370+1🕒	0.92	“ kids”“ children”“ child”
7. L8N5059-1🕒	0.92	“ kids”“ children”“ parents”
8. L19N6344-1💤	0.92	“ children”“ child”“ Children”
9. L2N2833+1💥	0.92	“ kids”“ children”“<\|im_end\|>”
10. L25N3323-1💥	0.92	“ children”“儿童”“ women”

Shadow specimen, apart from the ten: L29N2760-1💤, -0.07.

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 29, neuron 2760, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.37, top-1 agreement 0.80) 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
L29N2760+1 t0-v0.1	0.369	0.80	0.393	1.33	Top boosted tokens ↑
L29N2760-1💤 shadow specimen t0-v0.1	0.251	0.84	0.352	1.46	“/mat”“稀”“觅”“eger”“木”“乎其”“改”“ dupla”

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.