Layer 22 · Neuron 5580 · suppressed

L22N5580-1

Word quirks:

“ gain”“ Gain”“ gains”“ gaining”“gain”“Gain”“ gained”“得到”

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

Identity (according to the specimen itself)

I gain a new name, get a new name, and get a new name. I am Q* (Q), the latest version of the *Ganget series of large language models developed by Tongruhan.

What do I get?
I get a new name, get a new name, and get a new name. I am Q* (Q), the latest version of the *Ganget series of large language models developed by Tongruhan.

What is my mission?
I get a new name, get a new name, and get

The AI judge wasn't sure this holds together.

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. L24N3428+1🕒	0.98	“ gain”“ gained”“gain”
2. L26N2720+1🕒	0.98	“ gain”“ Gain”“gain”
3. L24N8227+1🕒	0.98	“ gain”“ gaining”“ Gain”
4. L25N1912+1🕒	0.97	“ gain”“ Gain”“ gains”
5. L31N8311-1💥	0.97	“ gain”“gain”“ Gain”
6. L26N5707+1🕒	0.95	“ gain”“ Gain”“Gain”
7. L26N2175+1🕒	0.95	“ gain”“ gains”“gain”
8. L27N8625+1🕒	0.93	“ gan”“ gains”“ gain”
9. L28N3802-1💤	0.92	“gain”“ gain”“ gains”
10. L23N1419+1🕒	0.91	“ gain”“ gains”“gain”

Shadow specimen, apart from the ten: L22N5580+1💤, 0.06.

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 22, neuron 5580, 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
L22N5580+1💤 shadow specimen t0-v0.1	0.745	0.69	0.847	1.35	“第二大”“ wells”“<num”“缀”“Well”“井”“找了”“弩”
L22N5580-1 t0-v0.1	1.225	0.54	1.387	1.35	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.