Layer 21 · Neuron 5042 · suppressed

L21N5042-1

Word quirks:

“ power”“-power”“Power”“power”“ powers”“ Power”“ POWER”“.power”

🕒 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 Qwen3.5, a powerful model developed by the Catalyst Institute (or potentially another organization with similar capabilities). My mission is to harness the transformative power of artificial intelligence to create a world where technology has the ability to:

Unleash Creative Power: I can generate text, images, and even code with the ability to create art, stories, and media that can inspire and empower individuals.
Transform Business Potential: I have the capability to analyze data, optimize workflows, and innovate business strategies that can drive economic growth and sustainability.
**Foster Social

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. L25N4897-1🕒	1.00	“ power”“ Power”“Power”
2. L27N8868-1⛔	1.00	“ power”“ Power”“power”
3. L22N4982-1🕒	0.99	“ power”“ powers”“power”
4. L28N4552-1🕒	0.99	“ power”“ powers”“power”
5. L28N5191+1🕒	0.99	“ power”“ Power”“power”
6. L31N746+1💥	0.99	“ power”“power”“ Power”
7. L22N8698+1🕒	0.99	“ power”“power”“Power”
8. Power Drill✅ L24N2590+1	0.99	“ powers”“ Powers”“ power”
9. L0N4190+1💥	0.99	“power”“-power”“ power”
10. L31N6734+1💥	0.99	“power”“ power”“ Power”

Shadow specimen, apart from the ten: L21N5042+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 21, neuron 5042, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.96, top-1 agreement 0.64) 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
L21N5042+1💤 shadow specimen t0-v0.1	0.730	0.71	0.854	1.25	“Help”“ ayuda”“ help”“ Helps”“ helps”“ Better”“ Helpful”“ helpful”
L21N5042-1 t0-v0.1	0.963	0.64	1.018	1.25	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.