Layer 9 · Neuron 3877 · suppressed

L9N3877-1

Word quirks:

“ ⏎”“ commonly”“ typically”“ ‘’”“ ——”“ ⏎⏎”“ similarly”“ particular”

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

Identity (according to the specimen itself)

I am an AI assistant, and my mission is to provide helpful, accurate, and tailored information based on specific goals or requirements. Whether it's solving a problem, creating content, or learning new skills, I aim to support you effectively. For example:

As a student, my mission might involve mastering a subject (e.g., studying for exams, understanding concepts).
As a professional, my mission could focus on achieving objectives (e.g., completing projects, meeting targets).
In personal growth, my mission might involve improving habits (e.g., fitness routines, dietary changes).
**In teamwork

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. L13N8498+1🕒	0.86	“ ⏎⏎”“ similarly”“ typically”
2. L14N7787-1🕒	0.80	“ ⏎”“ ⏎⏎”“ commonly”
3. L11N6995-1🕒	0.79	“ typically”“ usually”“ normally”
4. L8N1984-1🕒	0.79	“ usually”“ ‘”“ ——”
5. L11N4883+1🕒	0.79	“ commonly”“ ””“⏎”
6. L11N5188+1🕒	0.79	“ ⏎”“ typical”“ typically”
7. L10N894-1🕒	0.79	“ \"”“ “”“ similar”
8. L10N4168-1🕒	0.78	“ ‘”“ similar”“ usually”
9. L8N5198-1💤	0.78	“ commonly”“ standard”“ particular”
10. L10N2423-1💤	0.78	“<\|im_end\|>”“⏎”“ typically”

Shadow specimen, apart from the ten: L9N3877+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 9, neuron 3877, 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
L9N3877+1💤 shadow specimen t0-v0.1	1.000	0.66	1.228	1.22	“ ”“ '”“ physical”“ "”“ safety”“ simulation”“ ”“ "”
L9N3877-1 t0-v0.1	1.360	0.54	1.724	1.55	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.