INDEX

Explanations

words or phrases in all caps, and also finds decimals.

technical reports

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

 purpoſe

-0.74

 Monfieur

-0.73

 zubehör

-0.72

 fhew

-0.71

 Reſ

-0.69

 conflits

-0.69

 pleaſure

-0.69

Efq

-0.68

 Anſ

-0.67

 Majefty

-0.66

POSITIVE LOGITS

 defaultstate

0.48

zzleHttp

0.45

="@+

0.45

vorous

0.44

 estekak

0.43

kadot

0.43

GraphicsUnit

0.43

:][

0.41

@}

0.41

addCriterion

0.41

Activations Density 5.086%