INDEX

Explanations

multilingual context

This neuron consistently fires on words naming interface elements or spatial/layout terms—e.g. “окно,” “кнопка,” “границы,” “края,” “угол.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

をします

-1.11

 oamen

-1.05

 Männern

-1.02

 aislado

-1.00

を行いました

-1.00

 facendo

-0.99

To

-0.96

ための

-0.96

 ordentlich

-0.94

 tratto

-0.92

POSITIVE LOGITS

of

4.09

 оригіналу

1.38

 Кубка

1.35

ของ

1.23

 consé

1.17

국의

1.16

 juſ

1.16

 gerai

1.15

 illustre

1.11

1.10

Activations Density 0.019%