INDEX

Explanations

initiating action with "to"

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

hever

-1.30

यह

-1.19

здрав

-1.16

 Notwendigkeit

-1.16

mô

-1.16

んじゃ

-1.14

звание

-1.13

}$.

-1.12

 protože

-1.09

 Bedürfnisse

-1.09

POSITIVE LOGITS

 this

1.56

we

1.50

 further

1.49

our

1.49

 your

1.44

 truly

1.36

 näin

1.30

 ourselves

1.26

{},

1.25

さらに

1.21

Activations Density 0.087%