Layer 30 · Neuron 8303 · suppressed

L30N8303-1

Word quirks:

“ardi”“intis”“nitz”“nings”“kast”“天之”“koop”“uku”

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

Identity (according to the specimen itself)

I Qwen3.5 is a large-scale language model developed by Alibaba Cloud's Tongyi Lab. My primary mission is to empower users with intelligent, efficient, and ethically grounded assistance across diverse domains. Here's how I aim to deliver value:

Core Capabilities & Missions:

Universal Problem-Solving

From coding and data analysis to creative writing and logical reasoning, I strive to provide accurate, context-aware responses tailored to user needs.
Example: Debugging complex code, generating optimized algorithms, or distiling research papers into actionable

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. L30N6771-1🕒	0.69	“ki”“arca”“dech”
2. L26N2027-1💤	0.63	“kow”“endus”“.nasa”
3. L22N2057-1💤	0.63	“ixer”“日生”“esine”
4. L31N1472+1💥	0.63	“nok”“enco”“uera”
5. L31N4362+1💥	0.62	“namen”“ urat”“zana”
6. L31N6680-1💥	0.62	“anou”“ardy”“anko”
7. L23N3778+1💤	0.61	“ucu”“ardi”“ kard”
8. L21N5768+1💤	0.61	“ nik”“ipan”“4”
9. L30N8291-1💤	0.61	“arr”“strup”“zin”
10. L23N5536-1💤	0.61	“okon”“ikor”“kor”

Shadow specimen, apart from the ten: L30N8303+1🕒, 0.42.

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 30, neuron 8303, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.38, top-1 agreement 0.79) 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
L30N8303+1🕒 shadow specimen t0-v0.1	0.468	0.81	0.516	1.75	“els”“unter”“el”“ hisp”“塞尔”“upun”“oge”“very”
L30N8303-1 t0-v0.1	0.376	0.79	0.445	1.79	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.