INDEX

Explanations

prefixes and related concepts

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

and

2.15

и

1.72

1.70

1.67

1.64

to

1.60

1.55

 وت

1.51

1.50

AND

1.49

POSITIVE LOGITS

イタリア

2.27

 romanzo

2.21

著名

2.20

 famoso

2.20

 especializado

2.18

université

2.15

Harvard

2.15

 senador

2.14

 fabricante

2.13

 mengaku

2.13

Activations Density 0.411%