Layer 19 · Neuron 3067 · boosted

L19N3067+1

Word quirks:

“清洁”“ cleaner”“ clean”“ cleaning”“睡”

🕒 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 Qwen, a large language model developed by Doubai (Tencent). My mission is to provide high-quality, efficient, and safe services for users, helping them solve practical problems in areas like natural language understanding, code generation, data analysis, and more. I aim to offer clear, concise, and relevant responses while respecting user privacy and promoting fairness. How can I assist you today?

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. L27N9085+1🕒	0.94	“清洁”“ cleaning”“ cleaner”“ Cleaning”“Cleaning”
2. L24N7366-1🕒	0.94	“清洁”“ clean”“ cleaning”“干净的”“ Cleaning”
3. L19N5127-1💤	0.93	“ cleaning”“ cleaner”“ Cleaning”“ Clean”“清洁”
4. L25N3622-1🕒	0.93	“ cleaning”“清洁”“ clean”“ Cleaning”“clean”
5. L23N3023-1💤	0.93	“清洁”“ clean”“ cleaned”“ CLEAN”“ellent”
6. L30N819-1💥	0.92	“ clean”“ Clean”“clean”“Clean”“清洁”
7. L25N5677+1🕒	0.92	“ clean”“清洁”“ cleanliness”“干净”“Clean”
8. L20N4778-1🕒	0.92	“ cleans”“ cleaned”“ discharged”“ cleaner”“ clean”
9. L28N845+1🕒	0.91	“ clean”“ Clean”“clean”“ CLEAN”“-clean”
10. L28N1964-1🕒	0.91	“ clean”“ Clean”“Clean”“clean”“清洁”

Shadow specimen, apart from the ten: L19N3067-1💤, -0.01.

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 3067, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.67, top-1 agreement 0.68) 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
L19N3067+1 t0-v0.1	0.670	0.68	0.725	1.13	Top boosted tokens ↑
L19N3067-1💤 shadow specimen t0-v0.1	0.499	0.72	0.591	1.30	“ additions”“添加”“ adding”“ addition”“adding”

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.