INDEX

Explanations

programming variable and type names

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

 malgré

-0.97

 industrielle

-0.95

 має

-0.94

 السياس

-0.93

 ſhould

-0.87

惬

-0.87

 quinze

-0.85

趿

-0.84

尽快

-0.84

 يبدو

-0.84

POSITIVE LOGITS

 that

1.40

new

1.16

 first

1.10

 current

1.09

 like

1.07

 where

1.03

 their

1.00

 this

0.99

 "..\..\

0.94

 what

0.93

Activations Density 0.017%