INDEX

Explanations

following specific tokens

The neuron flags closing XML/HTML tags (i.e. tokens like “</example>” or “</ul>”).

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

to

-1.66

of

-1.36

in

-1.36

had

-1.33

can

-1.28

on

-1.26

 came

-1.24

 from

-1.23

at

-1.21

-1.20

POSITIVE LOGITS

 mété

1.37

にします

1.32

 baskılı

1.32

 spoko

1.27

 phénom

1.26

 mascotas

1.23

様な

1.20

 sécher

1.20

 istnieje

1.19

 größe

1.18

Activations Density 0.000%