INDEX

Explanations

types and formats

The neuron activates on numeric tokens or numeric expressions (numbers or measurements) in the text.

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

titial

-1.59

一些

-1.44

кое

-1.43

-1.40

ION

-1.39

</h3>

-1.37

這是

-1.37

я

-1.34

 для

-1.34

if

-1.34

POSITIVE LOGITS

礴

1.60

 вернулся

1.57

我了

1.55

lawesi

1.50

PEAT

1.50

呖

1.46

scribers

1.45

嚯

1.45

 maupun

1.44

碉

1.43

Activations Density 0.025%