INDEX

Explanations

contrast with the past

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

萼

-1.30

"));

-1.27

 meski

-1.26

Routledge

-1.24

and

-1.23

 mūsų

-1.22

theilung

-1.20

UNDP

-1.18

 हालांकि

-1.17

 although

-1.16

POSITIVE LOGITS

りました

1.56

it

1.48

 juist

1.27

!”

1.27

焢

1.25

same

1.24

越來

1.23

越來越

1.22

 vitesses

1.20

相对

1.18

Activations Density 0.067%