Layer 10 · Neuron 3842 · suppressed

L10N3842-1

Word quirks:

“ visually”“ visual”“ visuals”“ appearance”“ Visual”“ visibly”“ UI”“视觉”

🕒 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 a female name, but I can present the information in a way that might be easier to read if you prefer a different visual style. Here are a few suggestions for how you could interact with me:

As a text-based avatar (your preferred interface):

Option A: Use a standard on-screen font (e.g., bold, large) to help you see the information quickly.

Option B: Continue the conversation by selecting one of these options:

Ask your question using a customized font that matches your screen's display preferences.

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. L8N2367+1🕒	0.97	“ visual”“ visually”“ visuals”
2. L19N6689+1🕒	0.96	“ visual”“ visuals”“ visually”
3. L17N5254+1🕒	0.96	“ visual”“ visually”“ visuals”
4. L24N5292+1🕒	0.96	“ visual”“ Visual”“ visuals”
5. L17N5612+1💤	0.96	“ visuals”“ visually”“ visual”
6. L19N8952+1🕒	0.95	“ visual”“视觉”“ visuals”
7. L7N6596-1🕒	0.95	“ visual”“ visuals”“ visually”
8. L27N9056-1🕒	0.95	“ visual”“视觉”“ Visual”
9. L2N5644-1💥	0.95	“ visual”“视觉”“ Visual”
10. L11N8332+1🕒	0.95	“ visual”“ visually”“ visible”

Shadow specimen, apart from the ten: L10N3842+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 10, neuron 3842, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.99, top-1 agreement 0.46) 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
L10N3842+1💤 shadow specimen t0-v0.1	0.831	0.66	1.041	1.29	“⏎”“ ⏎”“ ‘”“ ⏎”“␍⏎”“‘”“ “”“⏎ ⏎”
L10N3842-1 t0-v0.1	1.993	0.46	2.392	2.11	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.