Layer 11 · Neuron 6092 · boosted

L11N6092+1

Word quirks:

“ here”“ now”“ typical”“ observable”“ general”“ specific”“ usually”“ 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 Qwen3.5, the latest large language model in the Qwen series. My mission is to provide intelligent, accurate, and helpful services across a wide range of tasks, from text generation to code development and data analysis.

Key aspects of my mission include:

Understanding human intent: Whether it's answering questions, generating content, or solving problems, I aim to bridge the gap between humans and machines.
Supporting diverse applications: From writing reports and creating software to analyzing data trends, I help users achieve their goals efficiently.
Ensuring reliability: By combining advanced reasoning with real

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. L9N3671-1💤	0.91	“ usually”“ typical”“ usual”
2. L8N2190-1🕒	0.90	“ usually”“ description”“ specific”
3. L8N7770+1🕒	0.90	“ typically”“ usually”“ typical”
4. L15N8891+1🕒	0.89	“ typical”“ typically”“ usually”
5. L11N2715+1🕒	0.89	“ typically”“ usually”“ here”
6. L9N5412+1💤	0.89	“ usually”“ typically”“ standard”
7. L9N6369+1💤	0.88	“ typical”“ usual”“ here”
8. L9N7191+1🕒	0.88	“ typically”“ usually”“ typical”
9. The Hand-Waver✅ L7N2792-1	0.88	“ usually”“ basically”“ now”
10. L8N1177-1💤	0.88	“ usually”“ typical”“ standard”

Shadow specimen, apart from the ten: L11N6092-1🕒, 0.34.

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 11, neuron 6092, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.61, 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
L11N6092+1 t0-v0.1	0.606	0.70	0.794	1.11	Top boosted tokens ↑
L11N6092-1🕒 shadow specimen t0-v0.1	0.894	0.65	1.131	1.31	“ already”“ yourself”“ similar”“ myself”“ exactly”“ created”“ similarly”“ exact”

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.