INDEX

Explanations

Greek letters, abbreviations, start of words

The neuron responds to Greek‐letter names (e.g. the names of fraternities/sororities or letter 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

很赞

-1.62

 ditt

-1.48

 AVEC

-1.45

让我

-1.43

 créer

-1.39

糴

-1.38

 ещё

-1.38

憷

-1.38

つた

-1.32

硌

-1.31

POSITIVE LOGITS

はこの

1.72

 setelah

1.63

急速

1.56

 spē

1.38

 автомо

1.37

isième

1.37

⃙

1.36

ميل

1.34

checkable

1.34

 tämä

1.34

Activations Density 0.008%