INDEX

Explanations

communication channels

The neuron strongly activates on tokens naming communication or contact methods—especially “email”/“e-mail,” “mail,” and “phone.”

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

 totalmente

-1.45

 vật

-1.35

 types

-1.35

you

-1.33

tht

-1.30

 WISCONSIN

-1.30

伫

-1.29

进入了

-1.28

 teď

-1.25

並んで

-1.25

POSITIVE LOGITS

or

1.74

。

1.38

𐄙

1.36

regelen

1.34

to

1.31

ledad

1.30

亊

1.30

precedented

1.27

iskt

1.22

 this

1.21

Activations Density 0.105%