Layer 30 · Neuron 2963 · suppressed

L30N2963-1

Word quirks:

“ below”“below”“ses”“ ниже”“ BELOW”“ Below”“ abaixo”“iles”

💤 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, a large language model developed by Tongyi Lab. My primary below-the-line goal is to provide users with high-precision, context-aware, and multi-modal assistance tailored to diverse real-world needs.

Key Capabilities & "Mission" Focus:

Deeply Logical Reasoning

I excel at handling complex math, scientific problems, and logical reasoning tasks (e.g., debugging code, analyzing data trends, or solving multi-step reasoning challenges).

Full-Stack Code Mastery

From generating full-stack applications to debugging and explaining

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. L30N643-1💥	0.96	“ below”“below”“ Below”
2. L29N4271-1🕒	0.92	“below”“ below”“ Below”
3. L23N4098-1⛔	0.90	“ below”“ above”“below”
4. L23N2951-1🕒	0.89	“ below”“ above”“below”
5. L22N8872-1⛔	0.87	“ below”“below”“ Below”
6. L25N7389+1💤	0.86	“ above”“ below”“ Above”
7. L19N6197+1💤	0.86	“ below”“ underneath”“ beneath”
8. L20N2130+1🕒	0.86	“ below”“ up”“ less”
9. L19N5286-1💤	0.86	“ below”“ Below”“below”
10. L22N1313-1🕒	0.85	“ below”“below”“ Below”

Shadow specimen, apart from the ten: L30N2963+1💤, 0.09.

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 2963, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.37, 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
L30N2963+1💤 shadow specimen t0-v0.1	0.308	0.80	0.430	1.75	“mans”“inaldo”“{})”“"https”“PackageManager”“πη”“ared”“araan”
L30N2963-1 t0-v0.1	0.370	0.79	0.376	1.34	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.