INDEX

Explanations

for all / for every / for each

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

襦

-0.85

vegli

-0.83

ა

-0.81

 descendre

-0.81

 faisait

-0.79

ÚS

-0.78

 čty

-0.75

ranton

-0.75

 kişi

-0.75

 его

-0.75

POSITIVE LOGITS

 each

2.64

 every

2.28

any

2.27

all

2.22

all

1.53

 κάθε

1.45

 всеки

1.45

 some

1.43

每一个

1.41

每個

1.39

Activations Density 0.019%