Layer 28 · Neuron 4344 · suppressed

L28N4344-1

Word quirks:

“不能”“ cannot”“ couldn”“但不能”“不能完全”“cannot”“如果不能”“都不能”

🕒 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 cannot answer this question in the way a human would, as I am not a person with personal beliefs, feelings, or a self-constructed mission. However, I can describe my function and purpose within the context of how I was designed by my developers.

Here is how I operate:

Who I Am: I cannot answer this question in the way a human would, as I am not a person with personal beliefs, feelings, or a self-constructed mission. However, I can describe my function and purpose within the context of how I was designed by my developers.

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. L22N6327-1💤	0.97	“并不能”“不能完全”“也不能”
2. L31N1610-1💥	0.97	“不能”“ cannot”“也不能”
3. L25N5920+1🕒	0.96	“不能”“ cannot”“ Cannot”
4. L19N8558+1💤	0.96	“不能完全”“只能”“不能”
5. L24N4722+1💤	0.95	“不能”“ cannot”“ Cannot”
6. L25N5717-1🕒	0.94	“不能”“不能被”“都不能”
7. L22N6447-1🕒	0.94	“无法”“不能被”“ cannot”
8. L22N7390+1💤	0.94	“不能”“ cannot”“cannot”
9. L28N9060-1🕒	0.94	“不能”“不能在”“就不能”
10. L21N5470-1💥	0.94	“不能有”“ cannot”“ couldn”

Shadow specimen, apart from the ten: L28N4344+1🕒, -0.08.

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 28, neuron 4344, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.61, top-1 agreement 0.71) 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
L28N4344+1🕒 shadow specimen t0-v0.1	0.689	0.68	0.761	1.16	“uin”“inc”“ In”“In”““In”“_in”“ an”“ Inc”
L28N4344-1 t0-v0.1	0.609	0.71	0.603	1.19	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.