INDEX

Explanations

is subjective and depends

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 substantiated

1.62

 同时

1.44

 timely

1.43

 toepassing

1.42

应用于

1.41

ărat

1.39

тивно

1.39

țiuni

1.38

不是

1.37

 Quindi

1.36

POSITIVE LOGITS

 resto

1.53

 concentrating

1.35

 gals

1.33

private

1.32

ier

1.32

 miscellaneous

1.31

剩下

1.30

眼鏡

1.28

xan

1.27

ອກ

1.27

Activations Density 0.241%