INDEX

Explanations

technical terms or code snippets

This neuron activates on Java import path segments naming classes or packages (e.g. “http,” “TextView,” “ListView”).

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

in

-1.59

-1.52

to

-1.32

on

-1.27

 más

-1.25

for

-1.23

 also

-1.20

 additional

-1.20

are

-1.17

POSITIVE LOGITS

Lø

1.55

 Änderung

1.32

ほら

1.30

をしました

1.29

acyjny

1.26

 كيفية

1.23

Ꭾ

1.20

が可能

1.20

謾

1.18

ră

1.16

Activations Density 0.000%