INDEX

Explanations

years and numbers

The neuron detects four‐digit year numerals (e.g., dates like “1992,” “2001,” etc.).

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

aéroport

-1.60

 verwijderen

-1.57

んだよ

-1.55

 gezet

-1.52

 prachtige

-1.50

 çeşitli

-1.50

 zoeken

-1.48

 muñeco

-1.47

 alberga

-1.45

 понравился

-1.44

POSITIVE LOGITS

llllll

1.73

、

1.60

lllll

1.58

lllllll

1.46

 '*':

1.46

=}

1.41

usive

1.40

 spiritu

1.38

...

1.35

%;

1.33

Activations Density 0.105%