Layer 18 · Neuron 7908 · boosted

L18N7908+1

Word quirks:

“ two”“ Two”“two”“ TWO”“ another”“Two”“ deux”“-two”

💤 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 and Qwen3.5, an AI assistant developed by Tongyi Lab. My primary mission is to empower users to achieve their goals through advanced AI capabilities. Here's how I support you:

Key Focus Areas:

Enhanced Capabilities:

Advanced Reasoning: Solve complex problems in math, logic, and scientific tasks with improved accuracy.
Multilingual Support: Assist users in two languages and regions, enabling global collaboration.
Content Creation: Generate high-quality text, including stories, reports, and creative

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. L20N3835+1💥	0.99	“ two”“ deux”“-two”
2. L24N5584-1🕒	0.98	“ two”“ Two”“Two”
3. L28N2187-1🕒	0.98	“ two”“two”“ Two”
4. Two Much✅ L23N4718+1	0.98	“ two”“ TWO”“Two”
5. L21N6009-1🕒	0.98	“ two”“ Two”“Two”
6. L24N6205+1🕒	0.98	“ two”“ Two”“two”
7. L25N4686-1🕒	0.98	“ two”“ Two”“Two”
8. L14N8508-1🕒	0.98	“ two”“ TWO”“two”
9. L26N6016-1💤	0.97	“ Two”“-two”“Two”
10. L28N3347+1🕒	0.97	“ Two”“two”“Two”

Shadow specimen, apart from the ten: L18N7908-1💤, -0.09.

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 18, neuron 7908, 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 0.69) 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
L18N7908+1 t0-v0.1	0.611	0.69	0.758	1.28	Top boosted tokens ↑
L18N7908-1💤 shadow specimen t0-v0.1	0.602	0.74	0.778	1.22	“ name”“ naming”“ balance”“平衡”“全县”“ designation”“你的名字”“ Nome”

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.