INDEX

Explanations

Okay, let's

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

awner

0.85

;)

0.82

 suspiciously

0.81

 obscurity

0.80

 herzlich

0.80

 anonymously

0.78

 unrecognizable

0.77

 mysteriously

0.77

行列

0.76

ールド

0.76

POSITIVE LOGITS

 addition

1.66

sofar

1.55

 contrast

1.50

asmuch

1.50

fact

1.48

 fact

1.45

addition

1.45

herent

1.40

ferences

1.38

 order

1.33

Activations Density 0.754%