INDEX

Explanations

academic journal abbreviations

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

-1.44

of

-1.32

 From

-1.20

it

-1.15

 yakni

-1.10

-1.09

 Four

-1.06

 mỗi

-1.05

 your

-1.04

様な

-1.04

POSITIVE LOGITS

がありますが

1.66

 journal

1.39

 faciliter

1.30

 Paramètres

1.30

もあって

1.27

 linge

1.26

erobic

1.26

 gants

1.25

が上が

1.24

 rafra

1.23

Activations Density 0.023%