INDEX

Explanations

of Asia

The neuron fires on numeric tokens representing decimal values (floating-point numbers).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

"));

-1.83

<td>

-1.80

was

-1.65

 продолжа

-1.55

etc

-1.50

：“

-1.50

unteers

-1.49

but

-1.47

 time

-1.45

")

-1.43

POSITIVE LOGITS

our

1.92

 them

1.74

 作り方

1.73

besos

1.51

 весьма

1.49

 konsult

1.46

獃

1.44

 noemen

1.41

 morena

1.38

Besitos

1.36

Activations Density 0.080%