INDEX

Explanations

names after commas

The neuron activates on numeric tokens—particularly digit strings and numeric values (ages, measurements, statistics, etc.).

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

 varandra

-1.66

 Other

-1.65

</em>

-1.57

 otros

-1.55

 geralmente

-1.52

 männis

-1.51

 otras

-1.48

 skolan

-1.46

 toujours

-1.44

 zawsze

-1.41

POSITIVE LOGITS

 ‘‘

1.56

?”

1.54

";

1.52

hud

1.47

åde

1.45

 регионе

1.45

 bụ

1.41

 quà

1.40

﴾

1.38

"}}

1.38

Activations Density 0.053%