INDEX

Explanations

excessive punctuation and capitalization

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

There

-1.81

 with

-1.73

If

-1.58

とはいえ

-1.39

 However

-1.38

Some

-1.37

 With

-1.35

 There

-1.33

ツイート

-1.31

小花

-1.30

POSITIVE LOGITS

</h4>

1.78

↵

1.63

 they

1.57

</strong>

1.55

-</

1.55

 hunne

1.54

鼉

1.52

</h2>

1.52

⎔

1.51

</b>

1.49

Activations Density 0.015%