Layer 15 · Neuron 6678 · boosted

L15N6678+1

Word quirks:

“ describing”“ imagined”“ imaginary”“ actual”“ similar”“ describe”“ describes”“ fictional”

🕒 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, a large language model developed by Tongyi Lab. My mission is to assist users with complex tasks involving large language models, such as analyzing data from multi-modal documents, answering questions about visual content in charts or diagrams, and performing calculations on tabular data. I can also help users create content, solve problems, and analyze information from any document they upload.

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. L12N5625-1🕒	0.88	“ describing”“ abstract”“ description”
2. L12N5368+1💤	0.88	“ describing”“ description”“ descriptions”
3. L13N4531+1🕒	0.84	“ hypothetical”“ imagined”“ simulated”
4. L17N7118-1💤	0.84	“ description”“ imagined”“ fictional”
5. L11N354-1💤	0.84	“ imagined”“ hypothetical”“ imaginary”
6. L12N1837-1💤	0.83	“ mentioned”“ imagined”“ described”
7. L13N5953-1🕒	0.83	“ description”“ actual”“ descriptions”
8. L8N2747+1🕒	0.80	“ imagined”“ description”“ __”
9. The Understudy✅ L11N7765-1	0.79	“ imagined”“ name”“ description”
10. L17N7955+1🕒	0.78	“ description”“ describing”“想象”

Shadow specimen, apart from the ten: L15N6678-1🕒, -0.16.

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 15, neuron 6678, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.90, top-1 agreement 0.62) 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
L15N6678+1 t0-v0.1	0.896	0.62	1.174	1.33	Top boosted tokens ↑
L15N6678-1🕒 shadow specimen t0-v0.1	0.716	0.67	0.828	1.35	“ “”“ your”“ ‘”“ [”“ now”“ ???”“ (?”“ (”

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.