INDEX

Explanations

with

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

of

-0.82

st

-0.61

 headed

-0.56

 giustizia

-0.52

PreferredItem

-0.52

in

-0.51

 initComponents

-0.50

ettare

-0.48

სქოლიო

-0.48

mnt

-0.47

POSITIVE LOGITS

 purpoſe

0.69

PyExc

0.66

 AssemblyCompany

0.65

ToAction

0.64

Clik

0.57

 coö

0.57

RegistryLite

0.57

 itſelf

0.57

Datuak

0.56

 ſtate

0.56

Activations Density 0.070%