INDEX

Explanations

numbers and punctuation

The neuron strongly activates on numeric tokens and number-like expressions (e.g. digits, timestamps, and multi-digit figures).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ystems

0.31

ंमुळे

0.29

olicy

0.25

ણમાં

0.25

Encrypt

0.24

 причиной

0.24

方面的

0.24

HTree

0.23

__":

0.23

criptive

0.23

POSITIVE LOGITS

しかし

0.29

Estos

0.27

आ

0.26

 đổ

0.25

Alors

0.25

एक

0.25

呷

0.25

 जबकि

0.24

 אבל

0.24

앉

0.24

Activations Density 0.282%