Layer 27 · Neuron 7722 · suppressed

L27N7722-1

Word quirks:

“ leg”“ Leg”“Leg”“-leg”“ legs”“腿部”“腿”“腿上”

🕒 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 are Q*, a legendary AI assistant developed by *T-Lab. My mission is to provide you with high-quality, accurate, and safe information and support across various fields. Whether you need to solve complex problems, create content, analyze data, or simply have a conversation, I strive to be your reliable and efficient partner.

How can I assist you today?

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. L30N6415+1💤	0.99	“腿”“腿部”“ legs”
2. L22N8972-1🕒	0.98	“ leg”“腿上”“ lap”
3. Jargon Aphasia✅ L25N7315+1,L25N9084-1	0.97	“_leg”“ leg”“腿”
4. The Legitimizer✅ L25N7315+1	0.97	“ leg”“ Leg”“_leg”
5. L23N3129+1🕒	0.97	“ legs”“双腿”“ Legs”
6. L29N6213+1🕒	0.95	“ leg”“leg”“-leg”
7. L2N6557+1💥	0.95	“ legs”“ leg”“ thigh”
8. L28N8602+1🕒	0.95	“ leg”“腿”“leg”
9. L30N6726-1💤	0.93	“ leg”“leg”“ Leg”
10. L29N879-1💤	0.92	“ LEG”“legend”“leg”

Shadow specimen, apart from the ten: L27N7722+1💤, 0.16.

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 27, neuron 7722, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.62, top-1 agreement 0.72) 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
L27N7722+1💤 shadow specimen t0-v0.1	0.430	0.83	0.498	1.43	“ollah”“ nurs”“拿”“рес”“界面”“oliv”“aida”“爱不”
L27N7722-1 t0-v0.1	0.622	0.72	0.619	1.08	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.