INDEX

Explanations

Q: or Not or Pac

The neuron activates on tokens that are numeric literals (sequences of digits, including version numbers and standalone numbers).

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

 lixo

-1.38

''');

-1.34

 MANY

-1.34

 katholischen

-1.30

 colega

-1.27

 Stadtteil

-1.24

 THOSE

-1.23

 प्रोड

-1.21

 Había

-1.21

 doucement

-1.20

POSITIVE LOGITS

וּ

1.52

}_{

1.39

 бампер

1.30

༘

1.26

ַּ

1.15

ætte

1.13

ּוֹ

1.09

of

1.05

YMS

1.05

ección

1.04

Activations Density 0.006%