Layer 11 · Neuron 4501 · boosted

L11N4501+1

Word quirks:

“ ‘”“ protection”“ '”“ protecting”“ thus”“ cost”“ protected”“ protects”

💤 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 large language model developed by Alibaba Cloud. My mission is to provide high-quality AI assistance by leveraging my advanced capabilities.

As Qwen3.5, I offer enhanced performance in several key areas:

Superior Language Understanding: With my foundation on the Llama 3 series, I ensure efficient processing of complex tasks.
Precise Code Generation: As a coding assistant, I help developers write optimized code, reducing development time and costs.
Efficient Resource Management: By optimizing for lower energy consumption, I contribute to reduced carbon footprints through more efficient

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. L8N6861+1🕒	0.89	“ protection”“ protected”“ basically”
2. L16N318+1💤	0.88	“ protects”“ protected”“ protecting”
3. L5N7261+1🕒	0.85	“ protection”“ protect”“ protected”
4. L9N3837-1🕒	0.85	“ protection”“ protecting”“ protected”
5. L10N7791+1🕒	0.85	“ protected”“ protection”“ protections”
6. L2N2829-1💥	0.85	“ protection”“ protect”“<\|im_end\|>”
7. L8N3942+1🕒	0.85	“ protected”“ protecting”“ protection”
8. L9N655+1🕒	0.85	“ protected”“ protection”“ protective”
9. L9N8830+1💤	0.85	“ protection”“ protected”“ protections”
10. L5N8805+1🕒	0.85	“ protection”“ protective”“ protecting”

Shadow specimen, apart from the ten: L11N4501-1🕒, 0.16.

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 4501, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.79, 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
L11N4501+1 t0-v0.1	0.787	0.65	0.906	1.31	Top boosted tokens ↑
L11N4501-1🕒 shadow specimen t0-v0.1	1.118	0.60	1.408	1.23	“ dialogue”“ actually”“ truly”“ really”“ dialog”“ your”“ specifics”“ particular”

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.