INDEX

Explanations

having read or reviewed

The neuron detects first‐person or addressed declarations asserting that the writer (or reader) “has” or “have” read, reviewed, or fully understands some instructions, terms, rights, or responsibilities.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 różnych

-1.34

 כדי

-1.23

 různých

-1.21

蹕

-1.18

 byla

-1.18

 разных

-1.16

 fantasmas

-1.15

එ

-1.11

alagi

-1.10

 międzynarod

-1.10

POSITIVE LOGITS

 according

1.26

webElement

1.18

 because

1.17

 from

1.16

 روسیه

1.12

and

1.11

 berdasarkan

1.10

 despite

1.05

our

1.03

getColumnIndex

1.02

Activations Density 0.052%