INDEX

Explanations

the ability to practice framework

This neuron is highly sensitive to the definite article “the.”

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

trecht

-1.46

Romain

-1.35

园

-1.34

 heet

-1.31

喔

-1.30

的問題

-1.29

身體

-1.27

 momentan

-1.26

 история

-1.24

شوف

-1.23

POSITIVE LOGITS

you

1.78

 when

1.56

 quando

1.52

ichtete

1.46

<bos>

1.45

 generalmente

1.42

 geralmente

1.41

if

1.39

enciales

1.37

Activations Density 0.107%