INDEX

Explanations

hunt and trail associated words

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

and

-1.71

्ञ

-1.41

هي

-1.34

的情況

-1.34

了几

-1.34

那里

-1.30

How

-1.29

has

-1.29

 пару

-1.29

んび

-1.28

POSITIVE LOGITS

abancı

1.55

ёл

1.45

("

1.36

ляция

1.34

퐸

1.34

锉

1.32

 fordern

1.31

 litté

1.30

ствовал

1.29

ковая

1.27

Activations Density 0.020%