Layer 9 · Neuron 5919 · suppressed

L9N5919-1

Word quirks:

“ ______”“ ___”“ _____”“ __”“ ________”“ ____”“ ‘”“<|im_end|>”

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

Identity (according to the specimen itself)

I am an AI assistant designed to provide information, answer questions, and offer assistance. I can help you with a wide range of tasks, from writing essays, coding software, or creating visual art.

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. L15N3166-1🕒	1.00	“ __”“ ___”“ ____”
2. L10N2581-1🕒	1.00	“<\|im_end\|>”“ ____”“ ______”
3. L10N1297-1💤	0.98	“ __”“ _____”“<\|im_end\|>”
4. L15N8909+1🕒	0.98	“ ___”“ __”“ ____”
5. L11N754-1🕒	0.98	“ ____”“ ______”“ __”
6. L10N759+1💤	0.98	“ ____”“ __”“ ______”
7. L19N698-1💤	0.98	“ __”“ ____”“ ______”
8. L6N1400-1💤	0.98	“ ____”“ ______”“ _____”
9. L10N1372-1🕒	0.98	“<\|im_end\|>”“ ____”“ ______”
10. L10N5354-1🕒	0.98	“ ____”“ **“”“ __”

Shadow specimen, apart from the ten: L9N5919+1💤, -0.10.

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 9, neuron 5919, 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
L9N5919+1💤 shadow specimen t0-v0.1	0.956	0.62	1.238	1.44	“ complete”“ description”“ usually”“ always”“ typically”“ describing”“ my”“ generally”
L9N5919-1 t0-v0.1	1.348	0.51	1.638	1.79	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.