INDEX

Explanations

why it is the case

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

NullException

1.57

صدر

1.46

 হিসেবে

1.40

 residente

1.37

 quedado

1.36

 විසින්

1.36

)،

1.35

ichia

1.34

देशक

1.34

 casado

1.34

POSITIVE LOGITS

<unused716>

1.70

<unused157>

1.63

<unused920>

1.56

 THEY

1.55

<unused1732>

1.54

<unused718>

1.54

<unused1843>

1.52

<unused275>

1.50

<unused162>

1.47

<unused983>

1.47

Activations Density 0.072%