INDEX

Explanations

systems including methods

The neuron fires on adverbial or prepositional qualifiers that introduce examples or elaborations (e.g. “including,” “specifically,” “usually”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-1.73

-1.52

it

-1.43

is

-1.36

 zamiast

-1.34

Any

-1.31

any

-1.30

 ponieważ

-1.27

Any

-1.27

 already

-1.26

POSITIVE LOGITS

 turist

1.72



1.63

 talle

1.57

 meni

1.55

 kust

1.54

 diny

1.54

 princípios

1.53

 jule

1.53

 牛肉

1.52

 sula

1.49

Activations Density 0.229%