INDEX

Explanations

string inclusion 'in'

The neuron strongly fires on instances of the token “in,” i.e. it detects occurrences of the word “in.”

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

different

-1.59

various

-1.50

through

-1.41

The

-1.40

to

-1.39

ng

-1.35

-1.30

an

-1.30

３

-1.28

They

-1.27

POSITIVE LOGITS

 your

1.80

you

1.78

 this

1.66

or

1.41

in

1.36

なんですよ

1.36

なんですね

1.32

 ktore

1.28

んですね

1.24

and

1.22

Activations Density 0.001%