INDEX

Explanations

phrases starting with "A"

The neuron fires on module import statements—especially the “import … from …” pattern (notably the “from” keyword and its punctuation).

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

 each

-1.38

can

-1.31

 nasceu

-1.30

他已经

-1.27

他会

-1.26

 [...]

-1.26

 sembra

-1.26

 förslag

-1.25

 conhece

-1.23

ぞれ

-1.21

POSITIVE LOGITS

 achats

1.44

เรื่อง

1.36

 genoux

1.31

娛樂

1.30

提到

1.27

 victoires

1.24

картинка

1.23

 things

1.18

購入

1.15

 appels

1.15

Activations Density 0.011%