INDEX

Explanations

confusion

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 stället

-1.00

 jouets

-0.79

 mères

-0.79

 varandra

-0.78

 financières

-0.77

 flesta

-0.75

 publiques

-0.74

 preuves

-0.74

 prisonniers

-0.74

 cérami

-0.73

POSITIVE LOGITS

ness

1.30

ly

1.29

ized

0.92

ity

0.84

ities

0.83

LY

0.80

ally

0.76

ian

0.76

ization

0.76

nya

0.75

Activations Density 0.106%