INDEX

Explanations

the letter "w" at the beginning of a text

w

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

 rač

-0.49

 vklju

-0.47

 izdel

-0.47

 comigo

-0.47

 gezegd

-0.47

 enggak

-0.43

 tekem

-0.43

 uporab

-0.41

lju

-0.41

 habet

-0.40

POSITIVE LOGITS

TagMode

0.96

Diweddarwch

0.88

contentLoaded

0.87

PreferredItem

0.82

AsUp

0.81

0.77

 فريبيس

0.77

principalTable

0.76

ContentAsync

0.71

AndEndTag

0.71

Activations Density 0.220%