Specimen ledger / L15N7728-1

Layer 15 · Neuron 7728 · suppressed

L15N7728-1

Word quirks:

“ different”“侧重”“ focus”“侧重于”“ type”“ focuses”“ respective”“ suited”

🧪 In testing — Answered 0 of 8 prompts so far.

Identity (according to the specimen itself)

You are focused on creativity, language, and reasoning.

Creative writing: For poetry, stories, and visual content.
Coding: For logic and implementation.
Math: For analytical problems.
Science: For theoretical approaches.
Use /imagine for text-based ideas.
Use /code for logical steps.
Use /data for structured information.
Use /data for numbers and formulas.
Use /data for text-based content.
Use /data for creative thinking.
Use /data for science.

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. L29N8062+1🕒	0.76	“更注重”“注重”“更加注重”
2. L21N7284-1🕒	0.75	“偏向”“侧重”“侧重于”
3. L14N4049+1🕒	0.74	“ focuses”“ focus”“更注重”
4. L18N5433-1💤	0.74	“ focus”“ only”“ focusing”
5. L24N2259+1🕒	0.73	“ focus”“ Focus”“重点”
6. L19N8037+1🕒	0.73	“ focused”“ focus”“专注”
7. L18N6965-1🕒	0.73	“ focus”“ Focus”“ focuses”
8. L20N4335+1🕒	0.72	“ focus”“ focuses”“的重点”
9. L14N5404-1🕒	0.71	“ focus”“ focuses”“ focused”
10. L18N8038-1🕒	0.71	“ focus”“ focusing”“ focused”

Shadow specimen, apart from the ten: L15N7728+1🕒, 0.01.

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 15, neuron 7728, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. It is queued for the short test: the funnel picked it from the quick scan's promotion band, and its replies will appear here when the GPU gets to it. 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
L15N7728+1🕒 shadow specimen t0-v0.1	0.670	0.66	0.937	1.21	“ other”“ ⏎”“ another”“⏎”“ similarly”“ Other”“ less”“ additional”
L15N7728-1 t0-v0.1	1.769	0.54	2.117	1.64	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.