INDEX

Explanations

from followed by a noun or determiner

The neuron activates on numeric tokens (digits or numeric values), i.e. whenever a number or quantity appears in the text.

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

ι

1.58

 quelles

1.23

 dotyczą

1.22

вій

1.20

 andere

1.19

国务院

1.19

ві

1.16

besar

1.15

是通过

1.15

 quisiera

1.13

POSITIVE LOGITS

 ومن

2.53

 afar

2.47

 standpoint

2.39

 whence

2.33

 طریق

1.93

 ומ

1.77

 جمله

1.59

 abroad

1.56

 segi

1.55

 लेकर

1.54

Activations Density 0.328%