INDEX

Explanations

countries and companies

The neuron fires on capitalized named entities—brand names, product names, countries, organizations, etc.

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

on

-1.73

 jepang

-1.48

 alltid

-1.45

 minggu

-1.44

ﾈ

-1.43

 bahkan

-1.38

 пароль

-1.35

一辈子

-1.30

она

-1.30

 terakhir

-1.29

POSITIVE LOGITS

1.73

 zrobić

1.35

1.30

<tr>

1.30

and

1.27

essentiel

1.27

1.23

assemblée

1.23

1.20

Activations Density 0.135%