Layer 17 · Neuron 7103 · boosted

L17N7103+1

Word quirks:

“ already”“ previously”“ only”“ now”“ exploited”“⏎”“ respect”“ digitally”

🧪 In testing — Answered 0 of 67 prompts so far.

Identity (according to the specimen itself)

Answered — readability check pending

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. L12N5331-1💤	0.74	“ currently”“ now”“ ONLY”
2. L11N7106-1💤	0.73	“ now”“ separate”“ ONLY”
3. L16N47+1🕒	0.71	“⏎”“ now”“ fake”
4. L11N3096-1💤	0.71	“ now”“ previously”“ formerly”
5. L12N4857-1🕒	0.70	“ before”“ now”“ already”
6. L15N2596-1🕒	0.70	“ now”“ continued”“ only”
7. L14N8711-1💤	0.70	“ earlier”“ now”“ previously”
8. L10N6461+1🕒	0.69	“ now”“ previously”“ earlier”
9. L12N1238+1🕒	0.69	“ recently”“ currently”“ now”
10. L12N2294+1💤	0.69	“ already”“ new”“ existing”

Shadow specimen, apart from the ten: Show and Tell✅, 0.16.

Cousins · specimens sharing a neuron

Specimen	Word quirks	Steering
Show and Tell✅ Shadow specimen	“ school”“ kids”“ suitable”	L17N7103-1

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 7103, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the full 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
L17N7103+1 t0-v0.1	0.639	0.72	0.727	1.13	Top boosted tokens ↑
L17N7103-1✅ shadow specimen t0-v0.1	1.173	0.55	1.440	1.49	“ school”“ kids”“ suitable”“ toddler”“ schooling”“ preschool”“ educational”“ age”

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.