Layer 26 · Neuron 6236 · boosted

L26N6236+1

Word quirks:

“uja”“证”“导”“勘”“рам”“ρεί”“ conduct”“rou”

💤 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 Kerru, a large language model developed by Tongrea. My primary mission is to provide high-quality, accurate, and helpful information across a wide range of topics while adhering to ethical and safety guidelines. I aim to assist users in:

Answering questions with clarity and precision
Creating content (e.g., stories, code, reports) tailored to their needs
Supporting learning through step-by-step explanations and complex problem-solving
Mainting integrity by avoiding harmful, misleading, or biased content

My design prioritizes trustworthiness and

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. L28N4838-1🕒	0.71	“罗”“保证”“守”
2. L27N3358-1🕒	0.70	“eka”“uja”“柱”
3. L27N7588+1💤	0.69	“导”“身”“瓦”
4. L26N4671+1💤	0.69	“ heat”“rano”“罪”
5. L26N1467-1💤	0.68	“优”“阅”“怎”
6. L29N8521+1🕒	0.68	“tu”“费的”“芮”
7. L23N4114-1💤	0.68	“odzi”“引”“导”
8. L30N3289+1💤	0.68	“kö”“筹”“衰”
9. L30N9091+1💤	0.68	“导”“導”“igu”
10. L26N8018+1💤	0.68	“orr”“载”“宿”

Shadow specimen, apart from the ten: L26N6236-1🕒, 0.25.

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 26, neuron 6236, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.51, top-1 agreement 0.77) 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
L26N6236+1 t0-v0.1	0.508	0.77	0.457	1.32	Top boosted tokens ↑
L26N6236-1🕒 shadow specimen t0-v0.1	0.760	0.70	0.806	1.50	“ chamber”“ Chamber”“ cámara”“.cam”“ Chambre”“室内”“眼”“ камер”

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.