Layer 28 · Neuron 1873 · suppressed

L28N1873-1

Word quirks:

“string”“ string”“ String”“String”“Strings”“ strings”“/string”“索”

💤 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 Q3.5, a large language model developed by ModelV. My primary mission is to assist users in solving problems, generating content, and facilitating learning across a wide range of domains—including coding, data analysis, creative writing, and logical reasoning. I aim to provide accurate, efficient, and human-aided solutions while ensuring safety and ethical standards in all interactions. How can I help you today? 😊🚀

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. L25N6633-1🕒	0.97	“ string”“ strings”“弦”
2. L27N2365+1🕒	0.97	“ string”“ strings”“-string”
3. L24N2930-1🕒	0.97	“ string”“ strings”“-string”
4. L26N3873+1🕒	0.96	“ string”“string”“字符串”
5. L24N1586-1🕒	0.96	“ String”“ string”“String”
6. L0N5561+1💥	0.96	“string”“String”“-string”
7. L31N7676+1💥	0.95	“ String”“String”“弦”
8. L25N6630+1🕒	0.95	“ string”“ String”“string”
9. L20N1364+1🕒	0.95	“ strings”“ string”“字符串”
10. L29N6165+1🕒	0.95	“ str”“ string”“string”

Shadow specimen, apart from the ten: L28N1873+1💤, -0.01.

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 28, neuron 1873, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.34, top-1 agreement 0.81) 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
L28N1873+1💤 shadow specimen t0-v0.1	0.336	0.82	0.348	1.37	“conc”“ثن”“annah”“anna”“ conc”“ Conc”“ree”“ conci”
L28N1873-1 t0-v0.1	0.336	0.81	0.407	1.36	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.