INDEX

Explanations

article followed by description

The neuron activates on numeric measurements expressed as decimal (floating-point) values.

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

 butik

-1.38

 فلس

-1.38

"",

-1.37

ﮑ

-1.34

klart

-1.30

 Frankrijk

-1.29

dagogik

-1.29

стта

-1.29

("

-1.24

 parfü

-1.24

POSITIVE LOGITS

 there

1.91

 There

1.57

 only

1.40

There

1.31

 ऐसा

1.26

したもの

1.25

Cuáles

1.23

豈

1.22

 misma

1.21

</h3>

1.20

Activations Density 0.100%