Layer 23 · Neuron 5917 · suppressed

L23N5917-1

Word quirks:

“ club”“俱”“ Club”“ dance”“ Disco”“ disco”“ clubs”“Club”

🕒 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 Club House, a digital party system designed to help people connect, dance, and have fun on the dance floor. My mission is to create a vibrant, inclusive, and safe space where people can move their bodies, express themselves, and enjoy the music together.

If you're looking for a way to get out on the dance floor and feel the rhythm of the night, Club House is the place to be! 🎧💊👢

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. L31N5345-1💥	0.87	“俱”“ club”“ Club”
2. L27N9014-1🕒	0.87	“ club”“俱”“club”
3. L26N1137-1🕒	0.87	“ club”“俱”“ Club”
4. L26N5028-1🕒	0.87	“俱”“ club”“club”
5. L26N6994+1🕒	0.86	“俱”“ club”“ Club”
6. Club Sandwich✅ L28N8417+1	0.85	“ club”“俱”“ clubs”
7. L2N6703-1💥	0.81	“ club”“<\|im_end\|>”“俱”
8. L22N8556-1💤	0.72	“俱”“ away”“ concessions”
9. L1N5459+1💥	0.72	“ club”“<\|im_end\|>”“ group”
10. L19N8185+1💥	0.68	“ club”“Pop”“ Pop”

Shadow specimen, apart from the ten: L23N5917+1🕒, 0.20.

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 5917, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.91, 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
L23N5917+1🕒 shadow specimen t0-v0.1	0.845	0.69	0.971	1.34	“ seed”“psilon”“种子”“CS”“梦”“梦里”“ Seed”“ Content”
L23N5917-1 t0-v0.1	0.914	0.66	0.933	1.15	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.