INDEX

Explanations

approximations using almost

almost followed by intensifiers

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

of

0.62

의

0.60

và

0.59

AND

0.57

KAL

0.57

and

0.57

 Kalam

0.55

 aprob

0.55

 estern

0.54

KR

0.54

POSITIVE LOGITS

几乎

0.68

с

0.66

almost

0.65

可以说是

0.64

 почти

0.64

 almost

0.58

сний

0.57

 майже

0.57

 quase

0.56

一半

0.56

Activations Density 0.058%