Layer 26 · Neuron 5671 · boosted

L26N5671+1

Word quirks:

“狐”“缓和”“UTTON”“ erw”“xt”“jak”“亡”“angani”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I Q* is a large-scale language model developed by *Qwen (part of the Qwen series). My core mission is to empower users with precise, efficient, and ethical AI assistance across diverse scenarios. Here's how I align with this mission:

Core Responsibilities

Accurate Information Handling

Provide verified answers in domains like science, technology, and culture.
Avoid hallucinations by cross-referencing trusted sources when possible.

Ethical & Safe Interaction

Prioritize safety by avoiding sensitive topics (e.g., political

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. L28N6501+1💤	0.71	“西斯”“擦”“緩和”
2. L28N1079+1💤	0.69	“狐”“�”“ordo”
3. L30N7345+1💥	0.69	“熊”“狼”“ Fox”
4. L27N7862-1🕒	0.69	“托”“猴”“烦躁”
5. L30N5248+1🕒	0.68	“忽”“ира”“ราง”
6. L30N345+1🕒	0.68	“ Borgo”“鹅”“狐”
7. L27N1743-1💤	0.68	“危”“狐”“ lager”
8. L25N3386-1💤	0.68	“ cánh”“ Trocken”“aza”
9. L31N2697-1💥	0.68	“隈”“狐”“优势的”
10. L30N8503+1💤	0.68	“包”“呼”“irk”

Shadow specimen, apart from the ten: L26N5671-1🕒, 0.14.

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 26, neuron 5671, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.55, top-1 agreement 0.75) 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
L26N5671+1 t0-v0.1	0.546	0.75	0.560	1.28	Top boosted tokens ↑
L26N5671-1🕒 shadow specimen t0-v0.1	1.287	0.56	1.291	1.01	“健康”“ health”“ Health”“健康的”“health”“Health”“的健康”“-health”

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.