Layer 10 · Neuron 1301 · suppressed

L10N1301-1

Word quirks:

“ sometimes”“ usual”“ normally”“ usually”“ typical”“ typically”“ if”“ needed”

🕒 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 an AI assistant. If you're having trouble with your Wi-Fi connection, try turning it off and on again. This simple step can often resolve connectivity issues. Your mission is to ensure a smooth experience by following these steps:

Check Your Internet Connection: Make sure your internet is working properly.
Clear Your Browser Cache and Cookies: Sometimes, clearing your browser cache and cookies can help if you're experiencing technical difficulties.
Try a Different Network: If you're still having issues, try connecting to a different network.

If the issue persists, consider that your device may need a

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. L13N7534+1🕒	0.97	“ usually”“ typically”“ typical”
2. L15N3456-1🕒	0.95	“ usually”“ usual”“ typically”
3. L11N5451+1🕒	0.95	“ usually”“ usual”“ typically”
4. L10N8560-1🕒	0.95	“ usually”“ often”“ sometimes”
5. L12N5942-1💤	0.95	“ usually”“ usual”“ typically”
6. L10N4952+1🕒	0.95	“ usually”“ often”“ sometimes”
7. L10N553+1🕒	0.95	“ typically”“ usually”“ typical”
8. L10N4342-1🕒	0.95	“ typically”“ usually”“ often”
9. L13N3816-1🕒	0.95	“ usually”“ typically”“ often”
10. L10N2511+1🕒	0.95	“ normally”“ usually”“ typically”

Shadow specimen, apart from the ten: L10N1301+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 10, neuron 1301, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.77, top-1 agreement 0.49) 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
L10N1301+1💤 shadow specimen t0-v0.1	0.938	0.66	1.188	1.25	“ ”“ ”“ #”“ ”“!”“ #”“ ⏎”“!”
L10N1301-1 t0-v0.1	1.770	0.49	2.051	1.88	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.