Layer 18 · Neuron 6117 · suppressed

L18N6117-1

Word quirks:

“ roles”“ Roles”“ role”“roles”“Roles”“ Role”“role”“Role”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am Qwen32, the latest model with a new identity. I can serve as your assistant in any role, such as:

Data Analyst: Analyzing data from different dimensions to identify key insights.
Role Player: Taking on roles like customer service or technical support to solve problems.
Creative Partner: Generating content for marketing, design, or research projects.
AI Trainer: Providing training data for machine learning models.

My mission is to empower you by turning knowledge into actionable insights—ensuring that every interaction helps you achieve your goals. Let’s swap roles! �

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. L26N7259+1🕒	0.98	“ Role”“Role”“ role”
2. L27N4191-1🕒	0.97	“ role”“ roles”“ Role”
3. L18N7039+1🕒	0.96	“ role”“ roles”“的角色”
4. L2N4187+1💥	0.96	“<\|im_end\|>”“ roles”“ role”
5. L24N5930+1🕒	0.95	“ role”“ Roles”“角色的”
6. L22N6291+1🕒	0.95	“ role”“ Role”“Role”
7. L31N1773-1💥	0.95	“ role”“role”“ Role”
8. L22N5194-1🕒	0.95	“ role”“ roles”“Role”
9. L20N7518-1🕒	0.94	“ role”“role”“ Role”
10. L23N6250+1🕒	0.94	“ roles”“ titles”“ role”

Shadow specimen, apart from the ten: L18N6117+1💤, 0.08.

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 18, neuron 6117, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.80, top-1 agreement 0.67) 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
L18N6117+1💤 shadow specimen t0-v0.1	0.556	0.72	0.588	1.18	“imal”“imon”“ checks”“ Old”“ 띄”“ingham”“的理想”“ federal”
L18N6117-1 t0-v0.1	0.795	0.67	0.972	1.30	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.