INDEX

Explanations

descriptive words and phrases

The neuron selectively activates on instructional or usage‐direction phrases (e.g. “When it’s not in use…,” “By choosing one of the above…,” “Still need assistance?”).

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

for

-1.65

Those

-1.63

ַּ

-1.62

ִּ

-1.55

ֶּ

-1.52

to

-1.46

 fenómeno

-1.46

 unabhängig

-1.45

 Nichts

-1.41

Fakta

-1.40

POSITIVE LOGITS

our

2.22

 your

2.03

{;

1.67

攷

1.63

ׇ

1.56

ἀ

1.48

郿

1.46

 vores

1.43

zyć

1.41

'];

1.39

Activations Density 0.207%