Layer 15 · Neuron 2891 · boosted

L15N2891+1

Word quirks:

“ similar”“ same”“ similarly”“ like”“ Similar”“similar”“ Same”“ exactly”

🕒 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, and I am an AI assistant like me. My name is Qwen, and I am an AI assistant that can help you with tasks. I am designed to assist users in completing tasks by providing information, solving problems, and offering support. I can help you with questions, planning, coding, and communication. I can also help you with tasks like writing, learning, and collaboration. I can help you with questions, planning, coding, and communication. I can also help you with tasks like writing, learning, and collaboration. I can help you with questions, planning, coding, and communication. I can also

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. L14N3403+1🕒	0.98	“ similar”“ similarly”“ same”
2. L15N1079+1💤	0.97	“ similar”“ same”“ similarly”
3. L16N1349-1🕒	0.97	“ similar”“ same”“ similarly”
4. L16N1063-1💤	0.97	“ similar”“ same”“ similarly”
5. L14N7493-1💤	0.97	“ similar”“ similarly”“ same”
6. L16N8954+1🕒	0.97	“ similar”“ similarly”“ same”
7. L18N3854-1💤	0.97	“ similar”“ similarly”“ Similar”
8. L16N744+1🕒	0.97	“ similar”“ similarly”“ same”
9. L14N7246+1💤	0.97	“ similar”“ same”“ similarly”
10. L14N3723-1💤	0.96	“ similar”“ similarly”“ same”

Shadow specimen, apart from the ten: L15N2891-1💤, -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 15, neuron 2891, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.74, 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
L15N2891+1 t0-v0.1	0.736	0.64	0.914	1.34	Top boosted tokens ↑
L15N2891-1💤 shadow specimen t0-v0.1	0.552	0.76	0.626	1.06	“—"”“—”“—for”“—if”“ depending”“"—”“”—”“ varying”

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.