INDEX

Explanations

gauge what to change

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

 although

-1.45

 from

-1.42

 after

-1.38

 three

-1.32

 despite

-1.30

as

-1.29

 because

-1.20

 previous

-1.16

 予約

-1.12

 while

-1.11

POSITIVE LOGITS

 your

1.39

ässer

1.34

atorias

1.24

tieron

1.22

 ___________

1.21

תיים

1.20

PLICATION

1.17

1.16

onien

1.16

 nobre

1.13

Activations Density 0.277%