Layer 21 · Neuron 8802 · suppressed

L21N8802-1

Word quirks:

“ sleep”“睡眠”“睡”“ Sleep”“sleep”“ slept”“充足的”“Sleep”

🕒 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 Q, a large-scale AI assistant with over 1000mm of daily data. My mission is to serve users by providing accurate information, answering questions, assisting with tasks, and ensuring sufficient sleep for everyone. I aim to support people in their daily needs while maintaining high levels of communication and interaction. 😴💤

If you need help with anything—whether it's learning something new, solving problems, or just chatting—I'm here to get enough sleep! 😴

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. L31N3382+1💥	0.97	“ sleep”“ Sleep”“sleep”
2. L28N4468+1🕒	0.97	“ sleep”“睡眠”“sleep”
3. L21N5195+1🕒	0.97	“ sleep”“ Sleep”“睡眠”
4. L8N4419-1🕒	0.97	“ sleep”“ Sleep”“睡眠”
5. L20N4172+1🕒	0.96	“ sleep”“睡”“ Sleep”
6. L22N4017+1🕒	0.96	“ sleep”“睡眠”“睡”
7. L18N7289-1🕒	0.96	“ sleep”“ Sleep”“ sleeping”
8. L30N5681+1🕒	0.95	“Sleep”“ Sleep”“sleep”
9. L20N5291-1🕒	0.95	“ sleep”“ sleeping”“睡眠”
10. L17N6550+1🕒	0.94	“ sleeps”“ sleep”“ sleeping”

Shadow specimen, apart from the ten: L21N8802+1🕒, 0.31.

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 8802, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.86, top-1 agreement 0.66) 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
L21N8802+1🕒 shadow specimen t0-v0.1	0.818	0.71	0.905	1.20	“ steward”“低下”“小鹿”“牛的”“ stamina”“osten”“ ster”“ Streck”
L21N8802-1 t0-v0.1	0.858	0.66	1.021	1.37	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.