Layer 24 · Neuron 8503 · boosted

L24N8503+1

Word quirks:

“勋”“石柱”“ ukr”“博士后”“合物”“变”“ Prescription”“示”

💤 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 Kazara, a high-performance intelligent assistant designed to assist you with precision, efficiency, and adaptability. My core mission is to become your essential digital collaborator by:

Solving complex tasks – From analyzing data and writing code to crafting creative solutions, I handle challenges across technical, analytical, and creative domains.
Enhancing productivity – I streamline workflows, provide actionable insights, and adapt to your unique needs whether you're working on research, development, or daily tasks.
Ensuring quality & safety – Every response is rigorously checked for accuracy

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. L26N8298-1💤	0.83	“仕”“商”“坚毅”
2. L28N6339+1💤	0.83	“太”“芮”“玉”
3. L28N8031-1💤	0.83	“onan”“化”“硫”
4. L28N3288-1🕒	0.82	“维”“阳”“象”
5. L26N4977-1🕒	0.82	“磁”“三”“牺”
6. L27N8925-1🕒	0.82	“状”“向上”“季”
7. L30N404-1🕒	0.82	“ubar”“绎”“无忌”
8. L31N2270+1💥	0.82	“治”“色”“锡”
9. L31N6125+1💥	0.82	“晦”“德基”“利器”
10. L27N6399+1💤	0.82	“位”“绝对”“引”

Shadow specimen, apart from the ten: L24N8503-1💤, 0.17.

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 24, neuron 8503, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.73, top-1 agreement 0.75) 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
L24N8503+1 t0-v0.1	0.734	0.75	0.740	1.40	Top boosted tokens ↑
L24N8503-1💤 shadow specimen t0-v0.1	0.687	0.69	0.761	1.39	“固”“ solid”“Solid”“erein”“ uncompressed”“ Mau”“ Solid”“linger”

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.