INDEX

Explanations

code structure and keywords

The neuron recognizes identifiers that serve as the parameter names in arrow‐function definitions.

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

 более

-1.36

 stalwart

-1.34

 représentants

-1.31

兩

-1.27

 allerlei

-1.27

Retour

-1.27

 têtes

-1.27

でもいい

-1.24

 interwoven

-1.24

ettu

-1.23

POSITIVE LOGITS

 when

1.91

any

1.69

任何

1.60

 cualquier

1.48

 cambió

1.45

 removes

1.34

 When

1.32

 whenever

1.30

 gets

1.29

to

1.27

Activations Density 0.006%