Layer 21 · Neuron 612 · suppressed

L21N612-1

Word quirks:

“fuel”“燃料”“ fuel”“ Fuel”“Fuel”“取自”“ Wheel”“ Plen”

🕒 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 Q*, a large-scale data assistant developed by *Tian. My primary mission is to provide high-quality information, engage in fluent conversation, and assist with tasks such as content creation, analysis, and interactive dialogue. I aim to help users understand complex topics, solve problems, and enhance productivity through clear, accurate, and context-aware responses. 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. L26N9044+1🕒	0.94	“ Fuel”“ fuel”“Fuel”
2. L29N5073+1💥	0.94	“ fuel”“ Fuel”“Fuel”
3. L24N288+1🕒	0.93	“燃油”“ fuel”“ Fuel”
4. L28N3703-1🕒	0.93	“ fuel”“ Fuel”“燃料”
5. L23N5060-1🕒	0.92	“ fuel”“ oil”“油”
6. L2N6000-1💥	0.92	“ fuel”“ fuels”“ Fuel”
7. L6N8491-1🕒	0.92	“ fuel”“ diesel”“ Fuel”
8. L24N8276-1🕒	0.91	“ fuel”“ Fuel”“燃料”
9. L22N7184+1💤	0.91	“ fuel”“ fuels”“ Fuel”
10. L26N5876+1🕒	0.91	“ fuel”“燃油”“ Fuel”

Shadow specimen, apart from the ten: L21N612+1🕒, 0.21.

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 21, neuron 612, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.84, top-1 agreement 0.69) 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
L21N612+1🕒 shadow specimen t0-v0.1	0.888	0.70	0.954	1.42	“ finer”“ direction”“办事”“-gr”“细”“ebb”“方向”“ directions”
L21N612-1 t0-v0.1	0.839	0.69	0.956	1.18	Top boosted tokens ↑

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.