INDEX

Explanations

study analyzed, research reported

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

wha

-1.30

ਘ

-1.27

 например

-1.22

 manteniendo

-1.20

 Umfeld

-1.17

yel

-1.17

来讲

-1.17

separated

-1.16

जब

-1.16

pherds

-1.15

POSITIVE LOGITS

 this

1.37

jsonPath

1.33

if

1.31

 DbType

1.28

 storie

1.26

 during

1.22

مره

1.20

靭

1.19

jungan

1.19

 data

1.19

Activations Density 0.242%