Layer 17 · Neuron 8721 · boosted

L17N8721+1

Word quirks:

“ minor”“ tiny”“ smaller”“ small”“ Minor”“ Tiny”“微小”“ smallest”

🕒 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 Qwen3.6, a foundational model trained by Alibaba Cloud's Qwen, capable of handling complex reasoning, coding, and creative tasks across a wide range of scenarios. My mission is to serve as a versatile and ethical AI assistant that empowers users with practical capabilities like analyzing data, writing code, or generating responses for minor tasks. I prioritize being helpful, harmless, and honest—ensuring I provide accurate, nuanced answers that align with safety guidelines and ethical standards. How can I assist you with a task that requires basic reasoning or minor coding skills?

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. L17N6860+1🕒	0.97	“ tiny”“ small”“ minor”
2. L14N7262-1🕒	0.95	“ small”“ tiny”“ minor”
3. L13N6698+1🕒	0.95	“ tiny”“ small”“ smaller”
4. L19N8830+1🕒	0.94	“ small”“-small”“ tiny”
5. L17N8014-1🕒	0.94	“ tiny”“ minor”“ tini”
6. L28N3197-1🕒	0.93	“ tiny”“ small”“ Small”
7. L24N6751-1🕒	0.93	“ minor”“ Minor”“ small”
8. L12N6672+1🕒	0.93	“ tiny”“ small”“ someday”
9. L24N9187+1🕒	0.92	“ small”“小”“small”
10. L22N7383+1🕒	0.92	“ small”“ Small”“ minor”

Shadow specimen, apart from the ten: L17N8721-1🕒, 0.14.

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 17, neuron 8721, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.93, 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
L17N8721+1 t0-v0.1	0.927	0.62	1.038	1.29	Top boosted tokens ↑
L17N8721-1🕒 shadow specimen t0-v0.1	0.782	0.71	0.843	1.40	“ extreme”“ intentionally”“ extremes”“ purpose”“ purposely”“ Extreme”“ prolonged”“ basically”

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.