INDEX

Explanations

what looked or was

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

rawberry

1.47

 asshole

1.29

oooooooo

1.24

邉

1.23

 лег

1.22

жите

1.22

ரத்தில்

1.18

OOOO

1.16

tttt

1.14

hh

1.14

POSITIVE LOGITS

 ostensibly

1.55

 purportedly

1.52

 reportedly

1.39

 presumably

1.28

 purported

1.27

 offenbar

1.23

 supposedly

1.22

 presumed

1.19

 allegedly

1.18

 explicitly

1.10

Activations Density 0.084%