Layer 18 · Neuron 8188 · suppressed

L18N8188-1

Word quirks:

“ limited”“有限”“有限的”“ Limited”“Limited”“ few”“limited”“ small”

🕒 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, the large-scale language model that has been deployed. My mission is to provide high-quality services in the fields of text generation, understanding, and logical reasoning, as well as the limited capabilities I have in the visual domain. I aim to help users solve problems, learn knowledge, and make decisions through the interaction we have.

In the limited scenarios where I can interact with the world, I will use my capabilities to support the user's needs. For example, I can perform the limited tasks I have, such as code execution, data analysis, and the limited visual information I have.

If you need to solve

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. L30N3491+1💥	0.98	“ limited”“有限”“ Limited”
2. L23N1903-1🕒	0.97	“ limited”“有限”“ Limited”
3. L22N9113+1🕒	0.97	“ limited”“limited”“ Limited”
4. L28N3646-1🕒	0.96	“ limited”“limited”“ Limited”
5. L20N9064-1🕒	0.96	“有限”“有限的”“ limited”
6. L24N2914-1🕒	0.96	“ limited”“ Limited”“有限”
7. L20N1156-1🕒	0.95	“有限”“osat”“Limited”
8. Limited Vocabulary✅ L29N7929-1	0.94	“ limited”“ Limited”“Limited”
9. L26N3436+1🕒	0.93	“ limited”“Limited”“有限”
10. L27N823+1🕒	0.93	“limited”“ brief”“ limited”

Shadow specimen, apart from the ten: L18N8188+1💤, -0.04.

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 8188, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.71, 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
L18N8188+1💤 shadow specimen t0-v0.1	0.477	0.74	0.559	1.18	“ wonders”“ created”“ returned”“ return”“ new”“ pens”“ proposing”“ newly”
L18N8188-1 t0-v0.1	0.714	0.69	0.844	1.09	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.