INDEX

Explanations

very followed by adjective

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

 somewhat

-1.36

vább

-1.33

ícias

-1.28

 薬局

-1.25

 juſ

-1.23

somewhat

-1.23

 delightfully

-1.21

fillment

-1.20

より

-1.19

 lavabo

-1.19

POSITIVE LOGITS

 good

1.76

1.41

、

1.30

 important

1.27

 heavy

1.18

 high

1.18

 nice

1.16

 extensive

1.15

 strong

1.15

 well

1.15

Activations Density 0.041%