INDEX

Explanations

I would like to have

The neuron fires on numeric tokens (integers or decimals) 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

 they

-1.78

 двумя

-1.51

 köp

-1.46

 australiano

-1.43

 môžete

-1.42

tentu

-1.42

ﮑ

-1.41

 ועל

-1.40

sterreich

-1.37

它们的

-1.37

POSITIVE LOGITS

1.73

」、

1.66

to

1.62

穑

1.53



1.46



1.45

穡

1.44

珖

1.43

граммы

1.33

 актуаль

1.31

Activations Density 0.026%