Layer 23 · Neuron 1131 · boosted

L23N1131+1

Word quirks:

“ogn”“糜”“ Unused”“arka”“agung”“窒”“iegel”“ pys”

💤 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 Qwen, a large-scale spatial-spatial (spatial) multi-pooling language model developed by Tongya. My core mission is to rely on the power of data to provide high-quality, reliable, and diverse information services for users. I aim to:

Accurately understand and express complex concepts across various fields (e.g., science, humanities, technical domains).
Support diverse use cases like code generation, multi-language interaction, and multi-modal analysis.
Maintain ethical standards in all interactions, prioritizing user trust and

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. L31N6326-1💥	0.69	“稽”“苟”“prat”
2. L31N1968-1💥	0.69	“eses”“芮”“抽”
3. L28N7183-1💤	0.68	“indi”“mez”“�”
4. L28N5047+1🕒	0.67	“ẹt”“破”“akar”
5. L31N7891+1💥	0.67	“ogen”“劲”“iego”
6. L25N5930-1💤	0.67	“圩”“ubu”“eryl”
7. L31N4684-1💥	0.67	“屎”“ esplic”“urgi”
8. L31N1618+1💥	0.67	“卢”“Sel”“uves”
9. L27N2561-1🕒	0.67	“oss”“etre”“峻”
10. L30N5268-1💤	0.67	“ertz”“osti”“宗”

Shadow specimen, apart from the ten: L23N1131-1🕒, 0.06.

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 23, neuron 1131, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.62, 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
L23N1131+1 t0-v0.1	0.623	0.76	0.632	1.12	Top boosted tokens ↑
L23N1131-1🕒 shadow specimen t0-v0.1	1.500	0.58	1.617	1.71	“ standardized”“ Equal”“ equal”“标准化”“Standard”“标准化的”“ Standard”“ cleansing”

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.