INDEX

Explanations

of

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 referenties

-0.73

AndEndTag

-0.66

uxxxx

-0.59

principalTable

-0.53

 erklär

-0.51

hdysval

-0.51

ือง

-0.50

ake

-0.50

 engineering

-0.49

 сочи

-0.49

POSITIVE LOGITS

 CreateTagHelper

0.68

 serai

0.58

 Cutting

0.56

 cutting

0.54

Cutting

0.54

Cel

0.53

MENAFN

0.53

 ujednoznacz

0.51

cle

0.50

CPtr

0.50

Activations Density 0.001%