INDEX

Explanations

mathematical operations or code items

This neuron detects tokens that are names of code entities or API elements—i.e. class/type identifiers, register or field names, API method or resource names, and other CamelCase or all-caps tokens.

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

 tênis

-1.47

 insane

-1.38

 også

-1.36

 unicórnio

-1.35

 amazingly

-1.34

嵓

-1.34

 borboleta

-1.32

 incrível

-1.30

感じで

-1.29

 insanely

-1.29

POSITIVE LOGITS

 after

1.66

for

1.56

 from

1.52

in

1.52

if

1.45

 while

1.45

 provided

1.40

by

1.33

1.31

can

1.23

Activations Density 0.005%