Layer 24 · Neuron 2296 · suppressed

L24N2296-1

Word quirks:

“ proven”“ proved”“证明”“的可”“ prove”“ proves”“prove”“ prov”

🕒 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 a large network of the company (specifically, I am a series of the company that is a single-precise, single-procognitive, single-architecturable, single-philosophy, single-logic, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-structure, single-

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. The Proofreader✅ L25N1887-1	0.96	“ proven”“ proved”“ proof”
2. L30N6403-1💥	0.96	“ proven”“ proved”“ proves”
3. L24N5335+1🕒	0.95	“ prove”“证明”“ proves”
4. L20N3120-1🕒	0.94	“ proving”“证明”“ proof”
5. L21N5971-1💤	0.94	“证实”“证明”“ prove”
6. L20N8809-1🕒	0.94	“证明”“ proof”“ proving”
7. L27N1860+1💤	0.93	“ proved”“ prove”“proved”
8. L23N1902+1💤	0.93	“ proving”“证明”“证明了”
9. L23N7630-1💤	0.93	“证明”“ prove”“ proof”
10. L27N1381-1🕒	0.93	“证明”“ proved”“ proving”

Shadow specimen, apart from the ten: L24N2296+1💤, -0.05.

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 24, neuron 2296, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.69) 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
L24N2296+1💤 shadow specimen t0-v0.1	0.723	0.71	0.748	1.32	“agen”“onas”“ terrain”“ sak”“entas”“卷入”“ lager”“干了”
L24N2296-1 t0-v0.1	0.811	0.69	0.937	1.41	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.