INDEX

Explanations

cryptocurrency exchange

The neuron strongly fires on named entities and metadata in news‐style text—proper nouns (organizations, places), dates, and numeric expressions.

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

our

-1.87

ֺ

-1.61

my

-1.48

อร่อย

-1.42

 again

-1.34

ׇ

-1.33

工作

-1.30

lepiej

-1.29

 گیاه

-1.29

 растений

-1.28

POSITIVE LOGITS

🧫

1.68

וּ

1.52

ּוֹ

1.45

笑了笑

1.34

𓂂

1.30

 Tourisme

1.27

waża

1.27

 чехол

1.26

᾽

1.25

 veilig

1.23

Activations Density 0.023%