Layer 3 · Neuron 8925 · boosted

L3N8925+1

Word quirks:

“ steering”“舵”“ navigation”“ ship”“ direction”“ control”“ nav”“ directions”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am an artificial intelligence assistant, ready to assist you with any task or question. How can I help 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. L2N8104+1💥	0.78	“<\|im_end\|>”“ steering”“ controls”
2. L9N7773-1🕒	0.77	“ navigation”“航”“ navig”
3. L26N2585-1🕒	0.76	“ navig”“ navigation”“导航”
4. L6N3033-1💥	0.76	“ navigation”“ Navigation”“ breadcrumb”
5. L26N8850-1🕒	0.75	“ navigation”“导航”“ Navigation”
6. L23N6799+1🕒	0.75	“ navigation”“导航”“ Navigation”
7. L24N1846-1💤	0.75	“5”“导航”“Nav”
8. L18N7860-1🕒	0.75	“ navigation”“导航”“ navig”
9. L30N1654+1💥	0.74	“ navigation”“ Navigation”“ navig”
10. L3N4996+1🕒	0.74	“ navigation”“ navigate”“ navig”

Shadow specimen, apart from the ten: L3N8925-1🕒, 0.20.

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 3, neuron 8925, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L3N8925+1 t0-v0.1	2.025	0.51	2.437	1.87	Top boosted tokens ↑
L3N8925-1🕒 shadow specimen t0-v0.1	1.819	0.49	2.155	2.09	“ rugged”“ mount”“ mountain”“ terrain”“ -”“ mountains”“ rock”“ rocky”

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.