INDEX

Explanations

tech operations and multilingual data

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

 where

-1.61

 when

-1.60

all

-1.39

 using

-1.34

the

-1.31

 ricev

-1.31

 moglie

-1.27

to

-1.27

」。

-1.27

it

-1.26

POSITIVE LOGITS

 also

1.49

1.48

Также

1.45

ではでは

1.45

Also

1.42

もはや

1.37

 також

1.36

alions

1.30

 musíte

1.30

 THEOREM

1.29

Activations Density 0.108%