Layer 24 · Neuron 3446 · boosted

L24N3446+1

Word quirks:

“双”“ double”“ Double”“的双”“-du”“ dual”“ Dual”“雙”

💤 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 Neural (also known as Neuromin), a large-scale general-purpose AI model developed by Ningbo University. My primary mission is to provide high-quality, reliable, and efficient information across multiple domains, including but not limited to:

Scientific research: Supporting academic writing, data analysis, and technical problem-solving.
Engineering optimization: Offering solutions for coding, system design, and engineering challenges.
Educational support: Providing learning strategies, educational technology insights, and skill development guidance.
Productivity enhancement: Helping users manage tasks, write code

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. L26N2312+1🕒	0.99	“双”“ dual”“雙”
2. Doublespeak✅ L28N8233+1	0.97	“双”“雙”“ double”
3. L23N9031-1💤	0.97	“双”“ Double”“ double”
4. L31N2821+1💥	0.97	“ dual”“ Dual”“双”
5. L22N8581-1🕒	0.97	“ double”“的双”“双”
6. L26N3171+1🕒	0.97	“双”“雙”“的双”
7. L18N9094-1🕒	0.96	“ dual”“双”“ double”
8. L29N6872-1🕒	0.96	“双”“雙”“double”
9. L23N6622-1💤	0.96	“双”“ double”“双重”
10. L26N978-1🕒	0.96	“ double”“ Double”“double”

Shadow specimen, apart from the ten: L24N3446-1🕒, 0.12.

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 24, neuron 3446, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.69, top-1 agreement 0.76) 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
L24N3446+1 t0-v0.1	0.694	0.76	0.658	1.43	Top boosted tokens ↑
L24N3446-1🕒 shadow specimen t0-v0.1	0.813	0.68	0.955	1.28	“史”“預”“病史”“ historians”“History”“Been”“ history”“旭”

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.