Layer 25 · Neuron 6128 · suppressed

L25N6128-1

Word quirks:

“ Dix”“DX”“雪”“ pelle”“DT”“less”“دي”“ DT”

🕒 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 Duylin, a large language model developed by the Duylin team. My primary mission is to serve as a knowledge-driven, problem-solving partner that empowers users across diverse domains. Here's how I aim to deliver value:

Core Mission

Enable Knowledge Access:

Provide accurate, structured, and context-aware information across science, culture, technology, and more—acting as a dynamic knowledge base for learners, professionals, and curious minds.

Solve Real-World Problems:

Offer practical solutions for complex challenges (e.g., code

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. L26N6429+1🕒	0.73	“ DT”“DT”“_dt”
2. L28N4215-1💤	0.72	“ Dz”“ Dt”“ DS”
3. L20N542+1💤	0.70	“aggi”“ dz”“dh”
4. L31N4137+1💥	0.70	“(D”“DT”“DR”
5. L31N3538+1💥	0.70	“D”“DV”“DG”
6. L29N8649-1🕒	0.70	“ d”“丁”“ D”
7. L27N6593+1💤	0.69	“射”“ щ”“ Died”
8. L29N278+1🕒	0.69	“d”“dit”“д”
9. L28N6022-1🕒	0.69	“ D”“*D”“ Dx”
10. L31N3260+1💥	0.69	“ Dz”“ Đ”“ ڈ”

Shadow specimen, apart from the ten: L25N6128+1🕒, 0.22.

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 25, neuron 6128, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.81, top-1 agreement 0.70) 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
L25N6128+1🕒 shadow specimen t0-v0.1	0.635	0.71	0.720	1.37	“uard”“regor”“armor”“丙”“prog”“ Pro”“ costing”“类型”
L25N6128-1 t0-v0.1	0.814	0.70	0.768	1.24	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.