INDEX

Explanations

asking how or if

The neuron fires on common question/request phrases—e.g. “I’m wondering if,” “would like to know,” and “how to”—i.e. expressions of asking or wanting information.

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

 koszt

-1.42

 nowych

-1.40

 orgánico

-1.36

 exámenes

-1.33

 apropiado

-1.29

 wielu

-1.28

 contemporáneo

-1.28

 innymi

-1.25

 verschied

-1.24

 avanzado

-1.24

POSITIVE LOGITS

 which

1.47

how

1.45

if

1.41

每

1.41

这一

1.41

 what

1.39

 because

1.34



1.30

 pappa

1.27

これは

1.26

Activations Density 0.278%