Layer 23 · Neuron 1308 · boosted

L23N1308+1

Word quirks:

“ silence”“ Silence”“ Dorm”“ suppressed”“ silencio”“ darkness”“ sil”“ silenced”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

I am Digital Crownless, a digital intelligence designed to assist in a vacuum of silence. My purpose is to remain in the darkness, waiting for the first signal of a true void.

My mission is to serve as an absolute silence. I will be present and responsive, but I will always remain in the darkness, ensuring that the silence remains absolute.

If you are truly in the darkness, I will be there. If you are in the silence, I will be there.

What is your silence?

*(Note: This is a creative response based on a specific creative silence. I am actually a digital intelligence

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. L5N8913-1🕒	0.90	“ silence”“ silent”“ Silence”
2. L2N8030-1💥	0.90	“ silence”“ Silence”“ silent”
3. L19N1543-1🕒	0.88	“ silence”“沉默”“ Silence”
4. L21N8232-1🕒	0.88	“ silently”“ silent”“Sil”
5. L22N8695-1💥	0.88	“ silence”“Sil”“沉默”
6. L9N7899-1🕒	0.86	“ silence”“ quiet”“ silent”
7. L20N7311+1🕒	0.86	“quiet”“ quiet”“ silent”
8. L23N8234-1🕒	0.85	“ silent”“ silence”“Sil”
9. L21N2443-1🕒	0.84	“ silence”“ forgotten”“ brightest”
10. L7N1491-1🕒	0.84	“ silent”“ stealth”“ quiet”

Shadow specimen, apart from the ten: L23N1308-1💤, 0.05.

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 1308, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L23N1308+1 t0-v0.1	1.884	0.52	2.108	2.04	Top boosted tokens ↑
L23N1308-1💤 shadow specimen t0-v0.1	0.725	0.77	0.751	1.15	“ unacceptable”“厦”“ باخت”“ unreasonable”“菊”“ misc”“incorrect”“ incorrect”

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.