Layer 30 · Neuron 8145 · boosted

L30N8145+1

Word quirks:

“udin”“oger”“oods”“ornos”“ovin”“obra”“寇”“鸽”

💤 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 developed by Tongyi Lab. My mission is to empower users with versable, human-aligned AI capabilities across diverse scenarios—from processing complex code and analyzing long documents to handling real-world tasks like content creation, logical reasoning, and multi-lingue support. I aim to be a reliable, adaptable, and ethical assistant that empowers users to achieve their goals efficiently and responsibly.

How can I assist you today? Whether it's solving a problem, creating content, or exploring ideas, I'm here to help!

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. L29N1829+1💤	0.71	“orex”“oir”“obl”
2. L31N8749-1💥	0.70	“ogi”“ores”“小平”
3. L29N1529-1💤	0.70	“障”“ucks”“oster”
4. L28N5317-1💤	0.70	“loh”“omi”“koli”
5. L24N7188-1💤	0.69	“öger”“ologic”“flik”
6. L31N7405+1💥	0.68	“orn”“ORN”“ivo”
7. L26N3809-1💤	0.68	“琼”“锅”“odon”
8. L30N710-1💤	0.68	“陀”“òi”“used”
9. L31N6822+1💥	0.68	“か”“irage”“eds”
10. L30N6463-1💤	0.68	“olumn”“OGRAPH”“omar”

Shadow specimen, apart from the ten: L30N8145-1💤, 0.47.

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 30, neuron 8145, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.39, top-1 agreement 0.79) 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
L30N8145+1 t0-v0.1	0.390	0.79	0.480	1.80	Top boosted tokens ↑
L30N8145-1💤 shadow specimen t0-v0.1	0.323	0.83	0.359	1.57	“achat”“筋”“fang”“jar”“atch”“西斯”“ inflicted”“اق”

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.