INDEX

Explanations

the word "the"

the

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

Negative Logits

<bos>

-0.50

the

-0.50

-0.47

-0.45

-0.43

 delen

-0.42

’

-0.41

own

-0.40

 חיצוניים

-0.39

の

-0.38

POSITIVE LOGITS

клопе

0.68

 saites

0.61

 المعيارى

0.60

DockStyle

0.59

ыгана

0.59

cabulary

0.57

$_"

0.57

 intptr

0.57

 nahilalakip

0.56

دانشنامهٔ

0.55

Activations Density 2.813%