INDEX

Explanations

percent and %

This neuron fires on percentage figures—numeric values (often decimals) expressed as “percent.”

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

 emballage

-1.89

 composants

-1.75

 mété

-1.70

 expédi

-1.66

Они

-1.66

幀

-1.64

Mereka

-1.64

 Spéc

-1.63

 réfrig

-1.60

Sebagai

-1.59

POSITIVE LOGITS

in

1.99

to

1.80

 last

1.76

and

1.73

ти

1.54

end

1.47

The

1.46

it

1.37

on

1.34

1.32

Activations Density 0.012%