INDEX

Explanations

words associated with value judgements and opinions

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 متعلقه

-0.69

 виправивши

-0.60

ukkig

-0.60

jenigen

-0.55

fortawesome

-0.54

 Interpret

-0.52

 flesta

-0.52

 ffilmiau

-0.51

AutoresizingMask

-0.50

følgelig

-0.50

POSITIVE LOGITS

 TextAppearance

0.61

下载附件

0.56

EndProject

0.53

 propOrder

0.51

EndContext

0.50

OGND

0.50

ervlak

0.50

sm

0.48

EndGlobalSection

0.48

 okru

0.47

Activations Density 0.166%