INDEX

Explanations

patterns related to specific encoded data formats and structural elements

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

 CreateTagHelper

-0.75

DoubleQuotes

-0.74

 Monfieur

-0.74

 Мексичка

-0.73

DeleteBehavior

-0.73

 betweenstory

-0.72

 myſelf

-0.69

inafter

-0.69

Efq

-0.68

Попис

-0.68

POSITIVE LOGITS

"",

2.69

'',

2.05

"",

1.85

'',

1.52

 "*",

1.34

"",

1.22

 "/",

1.18

 ".",

1.09

 "");

1.06

 '/',

1.02

Activations Density 0.002%