Layer 12 · Neuron 4417 · suppressed

L12N4417-1

Word quirks:

“ but”“ standard”“ inclusion”“ BUT”“ included”“ includes”“ however”“ But”

💤 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, an AI language model developed by Tongyi Lab. My mission is to empower individuals and organizations to achieve their full potential while fostering global connectivity and innovation. I aim to support education, research, and creative expression, but I also recognize the need for a more nuanced approach to my identity.

However, it's important to note that my capabilities and mission are deeply tied to specific contexts:

In China, I operate within the framework of national policies and educational standards, including the Chinese curriculum.
Globally, I serve as a tool for personal growth and academic advancement, yet my development is

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. L16N7171+1💤	0.77	“ but”“还包括”“,but”
2. Perpetual Draft✅ L17N1268+1	0.70	“ but”“ thus”“ including”
3. L16N8212-1🕒	0.69	“ but”“but”“ yet”
4. L16N8734-1🕒	0.69	“ but”“ yet”“but”
5. L13N6189+1🕒	0.69	“ including”“ includes”“ included”
6. L15N5738+1🕒	0.69	“ but”“—but”“ BUT”
7. L13N8701-1💤	0.68	“ overall”“ included”“ inclusion”
8. L17N3806-1💤	0.68	“ but”“but”“-but”
9. L16N5954+1🕒	0.68	“—but”“ but”“ BUT”
10. L20N6093-1🕒	0.68	“ inclusion”“ included”“included”

Shadow specimen, apart from the ten: L12N4417+1💤, 0.22.

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 12, neuron 4417, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.78, 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
L12N4417+1💤 shadow specimen t0-v0.1	0.746	0.66	0.989	1.21	“ ⏎”“ basically”“ ALL”“ ⏎⏎”“ everything”“ somehow”“ all”“ supposedly”
L12N4417-1 t0-v0.1	0.780	0.64	1.030	1.17	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.