INDEX

Explanations

words and their suffixes

The neuron is a “rare‐word” detector, strongly activating on infrequently used, content‐bearing terms.

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

)};

-1.56

もちゃ

-1.45

 favoritos

-1.34

せます

-1.34

acetic

-1.34

下が

-1.30

 sustitu

-1.26

 personalizados

-1.25

 compuestos

-1.23

 Zoltán

-1.23

POSITIVE LOGITS

 also

1.77

的新

1.66

1.63

or

1.58

1.49

也

1.43

können

1.39

Ꮽ

1.39

Переваги

1.38

1.37

Activations Density 0.229%