INDEX

Explanations

susceptible to negative things

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

 påvir

-0.44

dymyr

-0.42

 AssemblyCulture

-0.41

iço

-0.41

Any

-0.40

 count

-0.40

prone

-0.40

ant

-0.40

-0.39

 susceptible

-0.39

POSITIVE LOGITS

findpost

0.88

ItemBackground

0.66

FunctionFlags

0.66



0.65

ImageContext

0.65

DoubleQuotes

0.65

 Wikimedijinoj

0.65

 Roskov

0.63

Manbalar

0.63

ValueStyle

0.63

Activations Density 0.006%