Layer 9 · Neuron 6890 · suppressed

L9N6890-1

Word quirks:

“ usually”“ directly”“ direct”“ larger”“ bigger”“ separate”“ description”“ typically”

💤 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 a large language model, typically referred to as an LLM. My primary function is to serve as a conversational AI that can process and generate text, code, and other complex data. This allows me to assist with tasks that would be too time-consuming or resource-intensive for a standard user to handle directly.

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. L10N5637+1🕒	0.80	“ separate”“ longer”“ usually”
2. L9N2094-1💤	0.79	“ typical”“ typically”“ usually”
3. The Layerist✅ L10N3050-1	0.77	“ directly”“ clearer”“ layered”
4. L11N5923-1🕒	0.75	“ normally”“ direct”“ directly”
5. L9N5282+1🕒	0.74	“ directly”“ general”“ typically”
6. L13N7063-1🕒	0.74	“ typical”“ generally”“ typically”
7. L10N7299+1🕒	0.74	“ typically”“ typical”“ normally”
8. L8N2519+1💤	0.74	“ typically”“ usually”“ typical”
9. L9N916+1💤	0.74	“ usually”“ typically”“ typical”
10. L10N8514-1🕒	0.74	“ basically”“ usually”“ typically”

Shadow specimen, apart from the ten: L9N6890+1🕒, 0.29.

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 6890, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.97, top-1 agreement 0.60) 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
L9N6890+1🕒 shadow specimen t0-v0.1	1.221	0.59	1.436	1.34	“ based”“ precision”“ precise”“ precisely”“based”“ Based”“ exact”“ accuracy”
L9N6890-1 t0-v0.1	0.970	0.60	1.235	1.45	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.