Layer 23 · Neuron 3143 · suppressed

L23N3143-1

Word quirks:

“等于”“ equality”“ equal”“equals”“equal”“_equal”“ Equal”“Equality”

🕒 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 QCEv2, a large language model developed by QCE (QinChu Energy). My mission is to provide high-quality, accurate, and helpful information across a wide range of domains, including but not limited to:

Energy solutions: Innovating in sustainable energy generation, storage, and transmission.
Scientific research: Supporting advancements in physics, chemistry, materials science, and engineering.
Global development: Contributing to sustainable growth, equality, and shared prosperity worldwide.
User empowerment: Enabling users to access reliable insights, solve complex problems, and make informed

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. L25N979-1🕒	0.99	“等于”“ equality”“ equals”
2. L22N4662+1💤	0.99	“ equal”“相等”“ Equal”
3. L22N1508+1💤	0.98	“ equal”“等于”“ equals”
4. L29N2550+1🕒	0.98	“ equality”“ равен”“ equal”
5. L22N1131+1🕒	0.97	“等于”“ equals”“ equal”
6. L22N1959+1🕒	0.97	“ equal”“等于”“ equals”
7. L31N5683-1💥	0.97	“ equal”“equal”“ Equal”
8. L21N3425+1💤	0.97	“ equal”“ equals”“等于”
9. L23N9143+1🕒	0.97	“ equals”“ equal”“ equality”
10. L0N6896-1💥	0.97	“ equal”“Equal”“ Equal”

Shadow specimen, apart from the ten: L23N3143+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 23, neuron 3143, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.90, top-1 agreement 0.68) 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
L23N3143+1💤 shadow specimen t0-v0.1	0.752	0.74	0.813	1.33	“-inline”“ block”“amen”“ Intr”“ Block”“中断”“ Cultura”“extr”
L23N3143-1 t0-v0.1	0.903	0.68	0.936	1.30	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.