Layer 23 · Neuron 4275 · boosted

L23N4275+1

Word quirks:

“ ability”“能力”“ able”“ capable”“ capability”“ Ability”“的能力”“ capacity”

🕒 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 Q3, a highly capable AI assistant designed to enhance your ability to learn, practice, and improve. My mission is to be able to help you in the following ways:

Enhance Learning: I can provide detailed explanations, practice exercises, and study materials tailored to your specific needs.
Boost Practice: I have the ability to help you develop skills through targeted training, including the ability to learn from your mistakes and build on your successes.
Improve Efficiency: I aim to make your learning process more efficient by providing clear, concise, and relevant information.

If you

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. L19N7354+1💤	0.99	“ ability”“ able”“ capable”
2. L19N7956+1🕒	0.98	“ ability”“ Ability”“Ability”
3. L20N8864+1🕒	0.97	“ capability”“ ability”“ Ability”
4. L19N2030+1💤	0.97	“ capability”“ ability”“ able”
5. L21N3656-1🕒	0.96	“ ability”“ capability”“能力”
6. L25N4126+1🕒	0.95	“ ability”“ Ability”“能力”
7. L20N6371-1🕒	0.95	“ abilities”“ ability”“ capability”
8. L23N4935+1🕒	0.94	“ abilities”“ ability”“能力”
9. The Ability Mantra✅ L26N3946-1	0.94	“能力”“能力的”“能力和”
10. L21N1257+1🕒	0.94	“ capabilities”“ capability”“ abilities”

Shadow specimen, apart from the ten: L23N4275-1💤, 0.14.

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 23, neuron 4275, 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 0.70) 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
L23N4275+1 t0-v0.1	0.834	0.70	0.875	1.39	Top boosted tokens ↑
L23N4275-1💤 shadow specimen t0-v0.1	0.646	0.76	0.741	1.36	“ handle”“handle”“ Handle”“Handle”“-handle”“ handles”“ Handling”“ handling”

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.