INDEX

Explanations

words that signal causality, reasons, or logical arguments

negations

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

Hauptartikel

-0.65

 cref

-0.58

WriteAttribute

-0.57

AndEndTag

-0.57

 EconPapers

-0.56

VersionUID

-0.54

Leírás

-0.50

öglichkeiten

-0.49

ItemBackground

-0.48

?>">

-0.48

POSITIVE LOGITS

not

2.70

not

2.31

Not

2.17

Not

2.06

NOT

1.84

NOT

1.83

 nicht

1.31

 niet

1.18

 не

1.16

 ikke

1.12

Activations Density 1.743%