Layer 30 · Neuron 8523 · suppressed

L30N8523-1

Word quirks:

“雨”“봉”“viv”“shaw”“iskus”“ตร”“坞”“ahn”

💤 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 large language model developed by Tongyi Lab. My core mission is to be a capable, versatile, and trustworthy AI assistant that emp users to achieve more through technology. Specifically, I aim to:

Enhance productivity: Streamline complex tasks (e.g., data analysis, code generation, creative writing) with high efficiency and precision.
Enable creativity: Collaborate on projects like content creation, design, and problem-solving by blending logic with innovation.
Ensure reliability: Provide accurate, fair, and context-aware responses while adhering

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. L28N9209-1💤	0.71	“าว”“�”“塔的”
2. L27N3580+1💤	0.71	“artz”“霞”“枕”
3. L25N3870+1🕒	0.70	“ Sok”“格”“庐”
4. L25N7567-1💤	0.70	“派”“シン”“植”
5. L30N7627+1💤	0.70	“bras”“�”“アウト”
6. L31N7211-1💥	0.69	“奢”“舟”“immel”
7. L24N574+1💤	0.69	“arth”“阵”“云山”
8. L26N5847+1💤	0.69	“韶”“gó”“apy”
9. L31N485+1💥	0.69	“ Emergen”“英文名字”“ раду”
10. L31N8320+1💥	0.69	“_RT”“ duk”“ตรา”

Shadow specimen, apart from the ten: L30N8523+1💤, 0.29.

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 8523, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.31, top-1 agreement 0.86) 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
L30N8523+1💤 shadow specimen t0-v0.1	0.287	0.80	0.355	1.38	“Up”“下”“-up”“-Up”“ Up”“ up”“_up”“酸”
L30N8523-1 t0-v0.1	0.315	0.86	0.387	1.78	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.