Layer 9 · Neuron 4426 · boosted

L9N4426+1

Word quirks:

“ representation”“ representing”“ representations”“ represented”“ represent”“ sim”“ depiction”“ represents”

🕒 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)

Hello! I'm a simple character with a very specific purpose.

I'm here to help you understand how I work. Here's my structure:

```
_ _
| | | |
| |_|_|
```

That's me! I'm ready for whatever comes next. 😊

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. L13N4761-1🕒	0.98	“ representation”“ representations”“ depiction”
2. L11N2682-1🕒	0.98	“ representing”“ represent”“ represents”
3. L10N1160+1🕒	0.98	“ representing”“ representation”“ represent”
4. L5N1227-1💥	0.98	“ representation”“ representing”“ represent”
5. L18N8315+1🕒	0.97	“ representing”“ representation”“ depiction”
6. L13N7123-1🕒	0.97	“ representing”“ represents”“ depicting”
7. L14N7853-1🕒	0.97	“ represents”“ representing”“ conceptual”
8. L10N4815+1💤	0.97	“ represented”“ representation”“ depicted”
9. L10N7585-1🕒	0.97	“ representation”“ representing”“ representations”
10. L10N7062-1🕒	0.96	“ representation”“ representations”“ representing”

Shadow specimen, apart from the ten: L9N4426-1🕒, 0.17.

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 9, neuron 4426, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 1.81, top-1 agreement 0.47) 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
L9N4426+1 t0-v0.1	1.812	0.47	2.075	1.80	Top boosted tokens ↑
L9N4426-1🕒 shadow specimen t0-v0.1	1.102	0.59	1.302	1.42	“ usually”“ completion”“ typically”“ identification”“ presentation”“ ____”“ seems”“ presumably”

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.