INDEX

Explanations

hold

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 suivantes

-0.56

ksessa

-0.52

 henkil

-0.51

 kyse

-0.50

 vectoriales

-0.50

berdayakan

-0.49

 jossa

-0.49

holdet

-0.48

setViewName

-0.48

 hold

-0.48

POSITIVE LOGITS

تقاوى

0.71

 Signalez

0.59

ISupport

0.57

atsun

0.56

 Universe

0.54

 backs

0.54

abestanden

0.53

 back

0.53

 universe

0.52

universe

0.52

Activations Density 0.005%