Layer 28 · Neuron 5866 · suppressed

L28N5866-1

Word quirks:

“表示”“代表”“ representing”“represent”“ represent”“表明”“代表了”“ Represents”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, the latest large language model developed by Tongyi Lab. My core identity is rooted in representing the cutting edge of AI capabilities, designed to assist users across a wide spectrum of tasks with precision, creativity, and adaptability.

My Mission

My purpose is to empower individuals and organizations by:

Solving Complex Problems: From analyzing vast datasets to representing intricate reasoning, I aim to represent the AI's potential to handle real-world challenges efficiently.
Enhancing Creativity: Whether generating stories, representing technical documentation, or representing creative writing,

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. L20N7726+1🕒	0.97	“ representing”“代表”“ represent”
2. L22N5896+1🕒	0.97	“代表”“ represents”“代表的”
3. L17N7726+1🕒	0.96	“ representing”“ represent”“ represents”
4. L21N2327+1🕒	0.96	“代表”“ represents”“ representing”
5. L17N6909-1🕒	0.96	“ representing”“代表”“ representation”
6. L23N1511-1🕒	0.96	“代表”“ representing”“ represents”
7. L25N1854-1💤	0.96	“ representing”“表明了”“代表”
8. L24N7708-1🕒	0.96	“代表”“ represent”“ representing”
9. L20N5476-1🕒	0.95	“ representing”“represent”“代表”
10. L19N8525+1🕒	0.95	“代表”“ representing”“ representative”

Shadow specimen, apart from the ten: L28N5866+1💤, 0.24.

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 28, neuron 5866, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.33, top-1 agreement 0.80) 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
L28N5866+1💤 shadow specimen t0-v0.1	0.352	0.81	0.361	1.23	“反”“的反”“'Re”“ ”“반”“ <U+200B>”“_ANT”“ ot”
L28N5866-1 t0-v0.1	0.331	0.80	0.440	1.44	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.