Layer 28 · Neuron 5599 · boosted

L28N5599+1

Word quirks:

“atico”“ refs”“شري”“عی”“onen”“kap”“iteli”“双子”

💤 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 (Tongyi) Lab. My primary mission is to empower users with intelligent, reliable, and versatile AI capabilities across diverse domains. Here's how I aim to deliver value:

Core Mission Pillars

Universal Intelligence

Master 100+ languages and 256K context windows to handle complex tasks like long document analysis, multi-step reasoning, and real-time data synthesis.
Support full-stack development (e.g., generating functional 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. L31N8518+1💥	0.64	“engt”“iteli”“istent”
2. The Mantra✅ L18N6866-1,L26N463+1	0.62	“籍”“ cie”“Bits”
3. L22N7919-1💤	0.62	“寿”“斯蒂”“atika”
4. L31N719-1💥	0.61	“شارات”“alni”“äinen”
5. L24N3590+1💤	0.61	“ssp”“мили”“терна”
6. L30N3890+1💤	0.61	“elius”“alat”“ Mód”
7. L26N9031-1🕒	0.61	“skim”“zey”“itek”
8. L30N2069-1🕒	0.60	“illery”“ استف”“rame”
9. L25N6676-1🕒	0.60	“egin”“اعي”“克鲁”
10. L31N7623+1💥	0.60	“藏”“čit”“رائ”

Shadow specimen, apart from the ten: L28N5599-1🕒, 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 28, neuron 5599, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.26, top-1 agreement 0.84) 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
L28N5599+1 t0-v0.1	0.260	0.84	0.267	1.35	Top boosted tokens ↑
L28N5599-1🕒 shadow specimen t0-v0.1	1.579	0.53	1.596	1.07	“ed”“ sed”“ Sed”“ sedi”“SED”“ сед”“sed”“ séd”

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.