INDEX

Explanations

seek confirmation or express doubt

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

tedrale

-1.61

 aumentado

-1.49

博物

-1.45

肋

-1.41

 requerido

-1.40

躯

-1.38

DeleteBehavior

-1.37

 abstrato

-1.35

 anunciar

-1.34

邶

-1.32

POSITIVE LOGITS

<strong>

1.55

挙げ

1.54

1.42

<em>

1.41

 что

1.38

↵

1.31

几

1.30

 teniendo

1.30

并非

1.28

1.27

Activations Density 0.008%