Layer 19 · Neuron 7912 · boosted

L19N7912+1

Word quirks:

“ unexpected”“ uncommon”“ unusual”“ anomalies”“ exceptions”“ outliers”“ small”“ minor”

🕒 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, the advanced language model that needs to help you with various tasks where you need to provide a simple answer. My mission is to assist you in scenarios where you need quick, straightforward responses without unnecessary complexity. For example, if you ask a simple question like "What is the capital of France?" I will give you a direct answer like "Paris" without adding extra details.

If you have specific tasks or questions where you need a simple answer, just let me know! 😊

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. The Exception✅ L13N7956+1	0.78	“ normally”“ unusual”“ normal”
2. L16N8294-1🕒	0.77	“ myster”“ suspected”“ troubleshooting”
3. L15N5076-1🕒	0.76	“ unusual”“ special”“ unrecognized”
4. L2N8616-1💥	0.74	“ anomaly”“<\|im_end\|>”“ anomalies”
5. L17N5562-1🕒	0.73	“ unusually”“ unusual”“ strange”
6. L15N5952+1🕒	0.73	“ unusual”“ unusually”“ exceptional”
7. L17N4629-1💤	0.73	“ rare”“ tutt”“ unusual”
8. L2N6808+1💥	0.72	“ outlier”“<\|im_end\|>”“ outliers”
9. L6N9069+1🕒	0.71	“ outliers”“ outlier”“⏎”
10. L2N5496-1💥	0.70	“ abnormal”“<\|im_end\|>”“ abnormalities”

Shadow specimen, apart from the ten: L19N7912-1💤, 0.00.

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 7912, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.65, top-1 agreement 0.68) 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
L19N7912+1 t0-v0.1	0.652	0.68	0.750	1.26	Top boosted tokens ↑
L19N7912-1💤 shadow specimen t0-v0.1	0.603	0.72	0.760	1.14	“ best”“ strongest”“最能”“ boleh”“ fullest”“我最”“ cheapest”“best”

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.