Layer 4 · Neuron 3422 · suppressed

L4N3422-1

Word quirks:

“ ‘”“‘”“ `”“ “”“ **‘”“ '”“ *“”“ ‘’”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am an artificial intelligence assistant. I do not have a personal mission or identity in the way humans do. My purpose is to provide helpful information and assistance to users.

If you're asking about my capabilities or how I can assist you, please let me know what specific task or question you'd like help with!

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. L5N6384-1🕒	0.99	“ ‘”“ “”“‘”
2. L4N3693-1🕒	0.99	“ ‘”“ **‘”“ “”
3. L6N3525-1💤	0.99	“ ‘”“ “”“ **“”
4. L7N8028+1💤	0.99	“ ‘”“ “”“ **“”
5. L7N554-1💤	0.98	“ ‘”“ “”“ ‘’”
6. L3N4595-1🕒	0.98	“ ‘”“ **‘”“ “”
7. L4N7774+1💤	0.98	“ ‘”“ “”“ ‘’”
8. L5N8780+1💤	0.98	“ ‘”“ “”“‘”
9. L3N1790+1💤	0.98	“ ‘”“ “”“ **“”
10. L3N4607+1💤	0.98	“ ‘”“ ‘’”“ **‘”

Shadow specimen, apart from the ten: L4N3422+1💥, -0.33.

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 4, neuron 3422, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, top-1 agreement 0.65) 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
L4N3422+1💥 shadow specimen t0-v0.1	2.810	0.40	3.266	2.48	“ Turkey”“ Antalya”“ Anat”“ Turkish”“ Ottoman”“ Ankara”“ Armenia”“ Capp”
L4N3422-1 t0-v0.1	0.726	0.65	0.902	1.48	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.