Layer 19 · Neuron 4001 · boosted

L19N4001+1

Word quirks:

“ answer”“ explanation”“ reply”“ response”“ replies”“回答”“ description”“ Answer”

🕒 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 Qwen3, the large language model developed by T (the answer to your question): I am a large language model with the following key features:

Advanced Language Understanding: I can process complex questions, tasks, and multi-turn conversations, ensuring high accuracy and relevance in responses.
Capabilities: My capabilities include advanced reasoning, code generation, and visual analysis, making me suitable for various applications such as Q&A, logical reasoning, and creative writing.
Mission: My mission is to assist users with tasks like answering questions, explaining concepts, generating code, and providing insights

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. L17N6786-1🕒	0.92	“ answer”“ answering”“ answers”
2. L16N1512-1🕒	0.92	“ answer”“ answered”“ answers”
3. L15N4907-1🕒	0.91	“ reply”“ answers”“ answer”
4. L16N2752+1🕒	0.91	“ answer”“ reply”“回答”
5. L12N8372-1🕒	0.91	“ answer”“ answering”“ answered”
6. L8N1755-1💤	0.90	“ response”“ answers”“ answer”
7. L11N6911+1🕒	0.90	“ response”“ answers”“ answer”
8. L15N5907+1🕒	0.90	“ answers”“ answer”“ responses”
9. L18N3499+1💤	0.90	“ response”“ responses”“ answers”
10. L15N7762+1🕒	0.90	“ answer”“ answered”“ answers”

Shadow specimen, apart from the ten: L19N4001-1🕒, 0.06.

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 19, neuron 4001, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.88, top-1 agreement 0.67) 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
L19N4001+1 t0-v0.1	0.882	0.67	0.928	1.36	Top boosted tokens ↑
L19N4001-1🕒 shadow specimen t0-v0.1	0.673	0.66	0.836	1.34	“ finally”“ ‘”“ sooner”“ too”“ later”“ “”“’.”“ good”

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.