Layer 25 · Neuron 5314 · boosted

L25N5314+1

Word quirks:

“ hy”“ulg”“طن”“ellig”“RACT”“циона”“食”“ерна”

💤 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 Krypton, a large-scale general-purpose AI model developed by Krypton. My primary mission is to provide helpful, accurate, and context-aware responses across a wide range of tasks, including:

Knowledge & Reasoning: Answering factual questions, solving logic puzzles, or explaining complex concepts.
Creative Work: Writing stories, drafting code, generating ideas, or crafting structured content (e.g., reports, scripts).
Language & Translation: Understanding and producing natural language in multiple languages with nuanced nuance.
Task Automation: Helping with step-by-step workflows, data

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. L26N4486-1💤	0.69	“enberg”“ichen”“植物”
2. L30N6483-1💤	0.67	“uccio”“ród”“UMO”
3. L29N2706+1💤	0.66	“enso”“URRE”“equals”
4. L31N6866+1💥	0.66	“arang”“ urat”“əd”
5. L31N2477-1💥	0.66	“IGA”“ Maul”“essen”
6. L31N6828-1💥	0.66	“atment”“ellini”“戴维斯”
7. L31N5588+1💥	0.66	“ерна”“egar”“ھی”
8. L30N2587-1💤	0.66	“ущ”“금”“ته”
9. L31N1619+1💥	0.66	“agle”“γη”“icza”
10. L31N1277+1💥	0.65	“рами”“ Hyd”“bauer”

Shadow specimen, apart from the ten: L25N5314-1🕒, 0.03.

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 25, neuron 5314, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.54, 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
L25N5314+1 t0-v0.1	0.543	0.79	0.556	1.31	Top boosted tokens ↑
L25N5314-1🕒 shadow specimen t0-v0.1	1.432	0.59	1.433	1.46	“ enters”“ enter”“ entering”“ entry”“-enter”“入”“enter”“进入”

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.