INDEX

Explanations

in-place or in position

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

每个

-1.55

hidupan

-1.42

 tomto

-1.41

 roar

-1.32

 tiap

-1.28

no

-1.27

ない

-1.25

 coscienza

-1.25

 نهاية

-1.25

в

-1.23

POSITIVE LOGITS

,”

1.47

),

1.46

 неё

1.41

 improving

1.38

”,

1.36

},

1.33

romas

1.30

',

1.30

",

1.29

]]

1.25

Activations Density 0.016%