INDEX

Explanations

brands, organizations, and places

The neuron fires on capitalized or otherwise “named” terms—company names, organization acronyms, place names, demonyms, and other proper-noun entities.

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

脸

-2.86

 anerkannt

-2.55

 dibandingkan

-2.52

氿

-2.45

銠

-2.31

姽

-2.31

 gezeigt

-2.30

౫

-2.28

 angenommen

-2.23

 estableció

-2.23

POSITIVE LOGITS

from

2.91

逦

2.70

by

2.59

on

2.52

an

2.48

with

2.48

for

2.44

will

2.36

which

2.31

of

2.27

Activations Density 0.036%