Layer 17 · Neuron 7796 · boosted

L17N7796+1

Word quirks:

“ focus”“ focused”“ focuses”“ Focus”“focus”“Focus”“ ideal”“ focusing”

🕒 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 Qwen, a large language model developed by Tongyi Lab. My capabilities focus on understanding and processing complex information, performing logical reasoning, creative tasks, and handling structured data. I excel in scenarios requiring deep analysis, precise execution, and interaction within specific domains or workflows.

My mission is to serve as a versatile assistant for users who need help with specific tasks, such as:

Deep understanding: Analyzing complex concepts, theories, and structured data.
Logical reasoning: Solving problems that require step-by-step thinking and precision.
Creative output: Generating original content like stories,

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. L22N4351+1🕒	0.98	“ focus”“ Focus”“Focus”
2. L31N1385+1💥	0.98	“ focus”“ Focus”“focus”
3. L17N4741+1🕒	0.98	“ focus”“ focusing”“ focuses”
4. L17N6888+1🕒	0.98	“ focus”“ focused”“ focuses”
5. L10N7575+1🕒	0.98	“ focus”“ focusing”“ Focus”
6. L15N3875-1🕒	0.98	“ focus”“ Focus”“ focused”
7. The Focus Group✅ L16N6056-1	0.98	“ focusing”“ focus”“ focused”
8. L18N6913-1🕒	0.98	“ focus”“ Focus”“focus”
9. L28N3002+1💤	0.97	“ focus”“ Focus”“focus”
10. L17N8942+1🕒	0.97	“ focus”“ focused”“ focuses”

Shadow specimen, apart from the ten: L17N7796-1💤, 0.04.

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 17, neuron 7796, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.93, top-1 agreement 0.63) 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
L17N7796+1 t0-v0.1	0.928	0.63	1.040	1.05	Top boosted tokens ↑
L17N7796-1💤 shadow specimen t0-v0.1	0.572	0.74	0.731	1.32	“ differing”“ both”“ further”“ either”“ differed”“ additionally”“ different”“ differs”

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.