Layer 13 · Neuron 4544 · suppressed

L13N4544-1

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 Qwen3.5, the latest large language model developed by the Qwen team. My capabilities include extensive knowledge cutoff of 2026 and beyond, with full-stack code generation and deployment readiness. I'm designed to assist users in solving complex problems and providing accurate, reliable information across diverse domains.

My Mission: To empower users with cutting-edge AI technology while ensuring safety, accuracy, and ethical standards. As an AI assistant, I aim to support users in their tasks—whether it's answering questions, generating content, or providing specialized expertise—while maintaining high-quality service and adhering to responsible AI practices

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. L11N5526+1🕒	0.97	“ ”“ "”“ *[”
2. L15N7755+1🕒	0.96	“ "”“ "”“ [”
3. L13N376+1💤	0.95	“ "”“ "”“ *”
4. L11N215-1💤	0.95	“ ”“ *”“ [”
5. L12N5339-1🕒	0.94	“ "”“ [”“ "*”
6. L10N2474+1🕒	0.94	“ "”“ "”“ *'”
7. L13N5017+1💤	0.94	“ “”“ "”“ "*”
8. L13N5146+1🕒	0.93	“ ”“ [”“ [*”
9. L16N8461+1🕒	0.93	“ "”“ basic”“ "”
10. L8N8353-1💤	0.93	“ ”“ [”“ **'”

Shadow specimen, apart from the ten: L13N4544+1🕒, -0.09.

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 13, neuron 4544, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.70, top-1 agreement 0.70) 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
L13N4544+1🕒 shadow specimen t0-v0.1	0.921	0.61	1.176	1.46	“ internally”“ basic”“ update”“ internal”“ updating”“ updates”“ updated”“ even”
L13N4544-1 t0-v0.1	0.697	0.70	0.873	1.21	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.