Layer 12 · Neuron 167 · boosted

L12N167+1

Word quirks:

“ seems”“ here”“ similarly”“ isn”“ maybe”“ doesn”“ somehow”“ —”

🕒 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 Qwen1.5, the latest version of the Qwen series from Alibaba Cloud. My mission is to empower users with cutting-edge AI capabilities and provide them with high-quality assistance and insights. I strive to make AI more accessible and useful for everyone!

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. L11N4166+1🕒	0.90	“ doesn”“ actually”“ isn”
2. L13N227+1🕒	0.84	“<\|im_end\|>”“ here”“ ____”
3. L15N2683+1💤	0.84	“ doesn”“ seems”“ isn”
4. L12N4760-1🕒	0.83	“ actually”“ doesn”“ somehow”
5. L15N4102+1🕒	0.82	“ sometimes”“ might”“ seems”
6. L15N5989-1🕒	0.82	“ doesn”“ somehow”“ seems”
7. L14N1413-1💤	0.82	“<\|im_end\|>”“ isn”“ doesn”
8. L13N506+1💤	0.81	“ seems”“ actually”“ doesn”
9. L14N8237+1🕒	0.81	“ ⏎”“ seems”“ somehow”
10. L13N4627+1💤	0.81	“ actually”“ basically”“ wouldn”

Shadow specimen, apart from the ten: L12N167-1💤, 0.10.

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 12, neuron 167, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.99, top-1 agreement 0.59) 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
L12N167+1 t0-v0.1	0.991	0.59	1.244	1.37	Top boosted tokens ↑
L12N167-1💤 shadow specimen t0-v0.1	0.665	0.71	0.787	0.96	“⏎”“<\|im_end\|>”“ [”“⏎⏎”“?”“?\”“...”“!”

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.