INDEX

Explanations

Out if people day medical

phrases introducing or drawing attention to information, like "find out", "how can", "to say", or "we've got".

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

mäs

-1.54

Shouldn

-1.42

 този

-1.39

adalah

-1.38

 denna

-1.38

臵

-1.35

㍑

-1.35

 realmente

-1.34

 ถ้า

-1.33

fär

-1.31

POSITIVE LOGITS

 would

1.83

1.54

 until

1.49

}{

1.42

 polícia

1.36

 could

1.34

 will

1.32

全員

1.30

 technology

1.30

 måtte

1.30

Activations Density 0.182%