INDEX

Explanations

the word "high", sometimes also activating on additional words that follow "high"

high

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

 autorytatywna

-0.61

kloped

-0.61

LookAnd

-0.57

 onOptions

-0.57

 réus

-0.56

 ")";

-0.55

 تضيفلها

-0.54

حياتها

-0.52

 näm

-0.52

 للاسماء

-0.51

POSITIVE LOGITS

 quality

0.60

fin

0.58

Datuak

0.55

gridx

0.54

portál

0.53

Välislingid

0.52

Fin

0.51

 grain

0.51

TabIndex

0.50

complexContent

0.50

Activations Density 0.155%