INDEX

Explanations

every word or anything

This neuron detects universal quantifier phrases of duration, such as “every minute.”

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

 something

-1.37

 someone

-1.31

 your

-1.31

 Hän

-1.27

 information

-1.27

 Giving

-1.26

 some

-1.24

ämään

-1.23

 Informationen

-1.23

POSITIVE LOGITS

of

2.28

ція

1.49

 including

1.46

 olika

1.46

 fiscale

1.44

}$,

1.40

 montrent

1.33

}$.

1.31

Toluene

1.31

*{-

1.30

Activations Density 0.009%