INDEX

Explanations

movements and position

The neuron fires on numeric and measurement tokens (digits, stats, URLs, and other numeric references).

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

ּוֹ

-2.64

 السياس

-2.28

ִּ

-2.00

וּ

-1.98

ֶּ

-1.91

וֹ

-1.89

ַּ

-1.88

 бампер

-1.73

־ה

-1.64

it

-1.56

POSITIVE LOGITS

ׇ

1.98

ܐ

1.73

 الجلد

1.66

 блюдо

1.60

 Prefeitura

1.54

 شهرد

1.49

 szkoły

1.48

 بازیگر

1.46

𝐑

1.45

🩴

1.45

Activations Density 0.280%