Layer 29 · Neuron 2390 · suppressed

L29N2390-1

Word quirks:

“are”“ARE”“ Are”“Are”“ ARE”“ are”“都是”“_are”

🕒 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 Qwen3.5, the latest large language model developed by Alare (通义千问) in China. My mission are to:

Empareare human potential: By combining advanced reasoning, code generation, and multilingual support, I aim to assist users in solving complex problems, creating content, and learning new skills.
Bridge knowledge gaps: With my vast training data and deep contextual understanding, I help simplify difficult concepts across science, technology, culture, and daily life.
Drive innovation: Through capabilities like long-context analysis, autonomous agent planning, and

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. L29N1723+1🕒	0.98	“are”“Are”“ Are”
2. L27N1986-1🕒	0.98	“are”“ Are”“Are”
3. L28N2617-1🕒	0.98	“are”“ Are”“ are”
4. Loophead✅ L30N161-1	0.98	“ Are”“Are”“are”
5. L31N906-1💥	0.97	“are”“ Are”“Are”
6. L25N1552-1💤	0.97	“ are”“ Are”“are”
7. L25N2670-1🕒	0.97	“are”“ are”“ARE”
8. L24N641-1🕒	0.96	“ Are”“Are”“ are”
9. L29N3033-1🕒	0.95	“are”“ Are”“ are”
10. L0N3497-1💥	0.94	“<\|im_end\|>”“ Are”“ are”

Shadow specimen, apart from the ten: L29N2390+1🕒, 0.36.

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 29, neuron 2390, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.58, top-1 agreement 0.79) 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
L29N2390+1🕒 shadow specimen t0-v0.1	0.430	0.77	0.491	1.16	“ is”“ é”“ è”“是可以”“是没有”“ was”“是一直”“是一年”
L29N2390-1 t0-v0.1	0.577	0.79	0.609	1.26	Top boosted tokens ↑

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.