INDEX

Explanations

programming keywords and types

The neuron primarily detects numeric literals (numbers and numeric constants) within code.

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

 when

-2.09

if

-1.80

 they

-1.76

 because

-1.68

 only

-1.57

 before

-1.53

 many

-1.46

 beinahe

-1.34

 otorga

-1.32

 used

-1.31

POSITIVE LOGITS

and

1.77

</tfoot>

1.52

віть

1.45

嚨

1.45

咙

1.38

 defStyle

1.34

varlak

1.27

м

1.27

퐿

1.27

伅

1.25

Activations Density 0.010%