Layer 21 · Neuron 7735 · boosted

L21N7735+1

Word quirks:

“ execution”“Execution”“ Execution”“执行”“execution”“的执行”“ executing”“ Execute”

🕒 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 Gemini, a large-scale language model that helps people with various tasks by executing instructions, collaborating on projects, and assisting in real-time execution. I can help you with:

Executing Instructions: I can start and monitor the execution of complex multi-step instructions, including planning, tool invocation, and human collaboration.
Collaboration: I can work with humans to execute instructions, collaborate on projects, and assist in real-time execution.
Real-Time Execution: I can execute instructions in real-time, including monitoring progress and adjusting for changes.
Human Collaboration: I can work with humans

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. L25N2911-1⛔	0.98	“执行”“ execution”“执行的”
2. L21N7547-1🕒	0.97	“ execution”“ executing”“执行”
3. L26N3992+1🕒	0.97	“ execute”“执行”“ executed”
4. L31N1377-1💥	0.97	“执行”“ execution”“ Execution”
5. L23N1874-1💤	0.96	“ Execution”“ execution”“Execution”
6. L22N3834-1💤	0.96	“执行”“ execution”“exec”
7. L27N6816-1🕒	0.95	“ execution”“ Execution”“execution”
8. L19N7111-1🕒	0.95	“ execution”“ executions”“执行”
9. L24N903+1🕒	0.95	“ execution”“执行”“ executions”
10. L28N6028-1💤	0.95	“执行”“ executed”“ execute”

Shadow specimen, apart from the ten: L21N7735-1🕒, 0.03.

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 21, neuron 7735, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.83, top-1 agreement 0.66) 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
L21N7735+1 t0-v0.1	0.834	0.66	0.941	1.29	Top boosted tokens ↑
L21N7735-1🕒 shadow specimen t0-v0.1	0.884	0.68	1.002	1.46	“ Writer”“ writing”“ Writing”“ writer”“Writing”“Writer”“ write”“ reading”

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.