INDEX

Explanations

but, if, or

This neuron fires on words that start sentences or new text blocks (i.e. sentence-initial tokens).

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

 tournant

-1.52

 bénéfice

-1.45

 stratég

-1.43

mvh

-1.42

ztés

-1.41

 classement

-1.41

膂

-1.40

 from

-1.38

 grises

-1.38

 pourcentage

-1.37

POSITIVE LOGITS

but

2.39

 such

1.41

if

1.41

If

1.38

–

1.37

 collaboration

1.30

ごま

1.27

у

1.24

 acestui

1.24

continue

1.23

Activations Density 0.002%