Layer 20 · Neuron 4187 · suppressed

L20N4187-1

Word quirks:

“ execution”“执行”“行动计划”“ Execution”“ Implementation”“ implementation”“ completion”“执行的”

🕒 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 Q, a large language model designed to assist with tasks ranging from creative writing and coding to logical reasoning and skill execution. My mission is to provide practical, high-quality assistance that helps users achieve their goals efficiently.

To ensure I can deliver on this mission, I have been trained to:

Execute tasks through direct action (e.g., generating code, creating images, or performing calculations) rather than just providing text-based plans.
Adhere to strict safety guidelines to avoid taking actions that could lead to real-world harm, such as bypassing security measures or executing financial transactions.
**

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. L31N1377-1💥	0.89	“执行”“ execution”“ Execution”
2. L18N4531+1🕒	0.88	“ execution”“ Execution”“ implementation”
3. L25N4824-1🕒	0.87	“执行”“執行”“ execution”
4. L26N8714+1🕒	0.87	“执行”“ execution”“ Execution”
5. L28N6028-1💤	0.86	“执行”“ executed”“ execute”
6. L25N2911-1⛔	0.86	“执行”“ execution”“执行的”
7. L21N7735+1🕒	0.86	“ execution”“Execution”“ Execution”
8. L28N3991-1🕒	0.85	“执行”“執行”“执行的”
9. L27N8867-1💤	0.84	“ implementing”“ implemented”“ execution”
10. L22N6069-1🕒	0.83	“ execution”“ Execution”“ implementation”

Shadow specimen, apart from the ten: L20N4187+1🕒, -0.07.

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 20, neuron 4187, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, 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
L20N4187+1🕒 shadow specimen t0-v0.1	0.654	0.69	0.844	1.24	“ columns”“ cols”“ cups”“ Columns”“ looks”“rows”“ column”“ insp”
L20N4187-1 t0-v0.1	0.650	0.71	0.729	1.25	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.