INDEX

Explanations

the words "want" and "like"

want

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

SequentialGroup

-1.02

 ModelExpression

-0.92

 InputDecoration

-0.89

 >=",

-0.83

 defaultstate

-0.82

IUrlHelper

-0.82

AsUp

-0.81

MessageOf

-0.79

Efq

-0.79

 esternos

-0.75

POSITIVE LOGITS

geschäft

0.50

an

0.45

pos

0.41

wiat

0.40

Ử

0.39

 tais

0.38

guidance

0.38

 just

0.38

to

0.38

inker

0.37

Activations Density 0.476%