INDEX

Explanations

for (in many languages)

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

MergeFrom

-0.50

?}",

-0.48

Зноскі

-0.47

 Italijani

-0.46

 Protection

-0.45

 跳转至

-0.45

abank

-0.44

pidou

-0.44

ناء

-0.44

 Hygiene

-0.43

POSITIVE LOGITS

for

1.84

for

1.73

For

1.42

For

1.41

FOR

1.34

FOR

1.16

 для

1.15

 voor

0.99

 untuk

0.97

 für

0.96

Activations Density 0.008%