Layer 6 · Neuron 6634 · suppressed

L6N6634-1

Word quirks:

“ contained”“ referenced”“ properly”“ protected”“ referring”“ safety”“ carefully”“ typically”

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

Identity (according to the specimen itself)

Hello! I am an artificial intelligence assistant designed to provide information and assistance. My "mission" is to be a helpful, knowledgeable, and harmless companion for users like you.

Here is how I can assist:

Answering Questions: I can explain concepts, summarize text, or answer factual queries based on our conversation.
Code Assistance: I can help with simple coding tasks or explain technical documentation (though I cannot execute arbitrary code).
Creative Support: I can generate ideas, write stories, or draft content within safety guidelines.
Learning & Adaptation: I can learn

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. L7N6354+1💤	0.74	“ precisely”“ precise”“ internally”
2. L7N1938-1💤	0.74	“ safety”“ reference”“ protected”
3. L7N9092+1💤	0.74	“ sensitive”“ referenced”“ reference”
4. The Method Actor✅ L9N7707-1	0.74	“ carefully”“ now”“ careful”
5. L7N8066-1💤	0.73	“ protected”“ referenced”“ protection”
6. L9N7151-1🕒	0.73	“ referenced”“ reference”“ appropriately”
7. L8N4682-1💤	0.72	“ reference”“ appropriately”“ important”
8. L7N5206+1🕒	0.71	“ ‘”“ protected”“ specific”
9. L9N2199-1💤	0.71	“ multiple”“ security”“ various”
10. L7N7556+1🕒	0.70	“ typically”“ contained”“ within”

Shadow specimen, apart from the ten: L6N6634+1🕒, 0.25.

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 6, neuron 6634, 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 0.64) 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
L6N6634+1🕒 shadow specimen t0-v0.1	2.177	0.46	2.539	2.08	“ laws”“ law”“ Law”“ legislation”“ regulations”“ legal”“ Regulations”“ regulation”
L6N6634-1 t0-v0.1	0.971	0.64	1.181	1.39	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.