INDEX

Explanations

words starting with Und

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

 Even

-0.77

ḏ

-0.75

 marshes

-0.74

 Employee

-0.74

Sagittarius

-0.73

 employee

-0.72

羅

-0.71

 Catalog

-0.70

 bolster

-0.69

 rival

-0.68

POSITIVE LOGITS

 zwar

2.94

 sice

1.30

Und

1.27

und

1.22

und

1.20

Und

1.16

UND

1.12

 though

0.94

 certes

0.93

UND

0.93

Activations Density 0.017%