INDEX

Explanations

permission to use or appeal

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

 direta

-1.46

-1.45

laştır

-1.40

 womit

-1.39

forChild

-1.32

neous

-1.31

|[]

-1.30

estimable

-1.29

 feitos

-1.28

 mengucap

-1.28

POSITIVE LOGITS

of

2.08

 already

1.77

 between

1.55

 anderen

1.51

 이미

1.34

 כבר

1.30

 they

1.27

 IMPROVEMENT

1.23

 greater

1.23

the

1.23

Activations Density 0.101%