INDEX

Explanations

compile and compiled code

The neuron activates on technical terms referring to code evaluation or compilation (e.g. “compile,” “compiled,” “eval,” “evaluate,” “code”).

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

manera

-1.41

Cocina

-1.41

 ihren

-1.38

 madres

-1.30

目前的

-1.26

colta

-1.22

蹀

-1.21

ﹶ

-1.21

'';

-1.20

 بعضی

-1.20

POSITIVE LOGITS

1.37

}$

1.29

见到

1.25

mathrm

1.25

И

1.23

 шанс

1.22

yor

1.21

никами

1.17

 ardından

1.17

":

1.16

Activations Density 0.079%