INDEX

Explanations

countries and regions

The neuron fires on geographic place names (e.g. countries and regions).

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

 actually

-1.89

 only

-1.71

 usually

-1.67

two

-1.63

 also

-1.55

 mostly

-1.48

 always

-1.46

 first

-1.41

 often

-1.41

 arguably

-1.40

POSITIVE LOGITS

 Nome

1.58

')],

1.40

?>

1.40

to

1.33

venez

1.30

 esquer

1.30

皷

1.30

tean

1.29

')),

1.27

笔

1.27

Activations Density 0.031%