INDEX

Explanations

first position

The neuron responds to references to the first element or its position in a sequence (e.g. “first byte,” “first letter,” “first character”).

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

not

-1.46

',

-1.31

 represent

-1.30

หน้าน

-1.27

 other

-1.23

 number

-1.22

UL

-1.21

NOT

-1.19

β

-1.18

POSITIVE LOGITS

 each

1.41

 farine

1.40

!（

1.39

 tabac

1.31

Lila

1.30

澳門

1.30

лектро

1.29

 komunikasi

1.28

 decorativas

1.27

 disponíveis

1.26

Activations Density 0.108%