INDEX

Explanations

to

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 '.';

-0.46

 about

-0.44

 complete

-0.43

 {}".

-0.42

の話

-0.41

nloa

-0.40

Société

-0.40

 whole

-0.39

istus

-0.39

 successful

-0.38

POSITIVE LOGITS

ImageContext

0.86

 estekak

0.80

 nahilalakip

0.76

 propOrder

0.75

/*

0.72

ScopeManager

0.71

 Roskov

0.70

المناصب

0.69

WebVitals

0.68

DoubleQuotes

0.68

Activations Density 0.186%