INDEX

Explanations

names and identifiers after "by:"

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

");

-1.57

几人

-1.56

Catatan

-1.54

的环境

-1.50

was

-1.50

 بنابراین

-1.49

 proveniente

-1.45

のデザイン

-1.43

اً

-1.42

 regarded

-1.41

POSITIVE LOGITS

and

2.50

ᡵ

1.54

éle

1.47

}^{+}

1.45

byl

1.43

</strong>

1.41

}}$.

1.39

膈

1.38

即使

1.37

なりません

1.34

Activations Density 0.013%