Layer 13 · Neuron 4106 · boosted

L13N4106+1

Word quirks:

“ here”“ my”“ our”“ ourselves”“ new”“ now”“ myself”“ your”

🕒 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 am Qwen3.5. Here's how I can help you:

Advanced reasoning: Solve complex math problems or logic puzzles (e.g., "If a factory produces 100 widgets per day, how many widgets can be produced in 7 days?" → Answer: 100 widgets/day × 7 days = 700 widgets).
Code generation: Create Python scripts for data analysis or automation tasks.
Multilingual support: I speak over 100 languages, including English, Spanish, French, German, Japanese, Chinese, and more.

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. L11N832+1💤	0.94	“ my”“ myself”“ our”
2. L7N2305-1💤	0.93	“ my”“ myself”“ ours”
3. L8N624+1🕒	0.92	“ my”“ our”“ myself”
4. L14N1466-1💤	0.90	“ ours”“ ourselves”“ here”
5. L6N3947+1🕒	0.90	“ my”“ now”“ info”
6. L13N1969-1💤	0.90	“ our”“ your”“ here”
7. L6N4089+1🕒	0.90	“ my”“ here”“ now”
8. L8N6896+1💤	0.89	“ my”“ now”“ new”
9. L12N4677-1💤	0.89	“ here”“ my”“ now”
10. L16N8404-1🕒	0.89	“ myself”“ my”“ yourself”

Shadow specimen, apart from the ten: L13N4106-1💤, 0.02.

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 13, neuron 4106, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.80, top-1 agreement 0.64) 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
L13N4106+1 t0-v0.1	0.795	0.64	0.991	1.29	Top boosted tokens ↑
L13N4106-1💤 shadow specimen t0-v0.1	0.662	0.72	0.761	1.07	“ similarly”“ “”“ ‘”“ ⏎”“ ⏎⏎”“ ___”“ __”“ again”

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.