INDEX

Explanations

name of the

The neuron fires on tokens containing numbers (i.e. numerical values) 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

少し

-1.64

 توسط

-1.53

 tuve

-1.41

 لأن

-1.40

 ال

-1.38

teilung

-1.38

 یکی

-1.37

Sejarah

-1.37

 antics

-1.34

志愿

-1.34

POSITIVE LOGITS

 with

1.81

who

1.73

how

1.73

 FEBRUARY

1.44

 that

1.42

 which

1.38

 where

1.37

苌

1.36

for

1.36

 もの

1.34

Activations Density 0.054%