INDEX

Explanations

code snippets

closing braces or brackets at the end of code blocks.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 both

-1.70

not

-1.53

 две

-1.51

 beide

-1.45

 drei

-1.43

 venue

-1.39

ANDUM

-1.38

 beiden

-1.37

or

-1.35

ございました

-1.34

POSITIVE LOGITS

 porcelaine

1.51

Menurut

1.46

 préférence

1.43

紲

1.43

 esmeralda

1.42

of

1.41

 土産

1.41

 poire

1.40

氳

1.39

おすすめ

1.38

Activations Density 0.005%