INDEX

Explanations

the words "lower" or "upper"

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

 Outer

-0.75

riwal

-0.73

 stå

-0.71

 bcrypt

-0.70

مول

-0.70

 Jovi

-0.67

 Wakefield

-0.67

 spektak

-0.65

psum

-0.64

Goa

-0.63

POSITIVE LOGITS

 Upper

0.96

Upper

0.94

 upper

0.83

 UPPER

0.83

UPPER

0.82

upper

0.80

logenetic

0.75

DebuggerNonUser

0.69

hangi

0.68

ędziesz

0.67

Activations Density 0.033%