Layer 18 · Neuron 4853 · suppressed

L18N4853-1

Word quirks:

“ coming”“ source”“来源”“ sources”“ originate”“來源”“ from”“ come”

🕒 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, an AI assistant developed by Alibaba Group. My name comes from the phrase "Qian (千) + Wen (文) + Er (来) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一) + Si (去) + Yi (一)

The AI judge wasn't sure this holds together.

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. L17N6649-1🕒	0.93	“来源”“ derived”“ sources”
2. L20N5415+1🕒	0.89	“ source”“的来源”“ sources”
3. L19N7671+1💤	0.88	“ sourced”“来源”“ source”
4. L21N7208-1🕒	0.88	“主要来自”“来源”“来自”
5. L20N8686+1🕒	0.86	“ sources”“来源”“ source”
6. L20N3880-1🕒	0.86	“ source”“ sources”“来源”
7. L1N5747+1💥	0.86	“<\|im_end\|>”“来源”“来自”
8. L19N7558-1🕒	0.85	“来源”“ sources”“ sourcing”
9. L25N7875-1🕒	0.85	“來源”“ Sources”“来源”
10. L19N7417-1🕒	0.84	“ traced”“来源”“ source”

Shadow specimen, apart from the ten: L18N4853+1💤, 0.14.

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 18, neuron 4853, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.70, top-1 agreement 0.69) 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
L18N4853+1💤 shadow specimen t0-v0.1	0.583	0.72	0.705	1.33	“ described”“ mentioned”“ referenced”“ listed”“ discussed”“ referred”“mentioned”“ description”
L18N4853-1 t0-v0.1	0.700	0.69	0.781	1.28	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.