INDEX

Explanations

Windows 10 versions

The neuron fires on the character “1” (especially when it appears in version‐number or numeric contexts).

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

 during

-2.25

’

-2.14

 sooooo

-2.14

瑧

-1.95

Despite

-1.94

Упаковка

-1.92

for

-1.91

 despite

-1.82

one

-1.79

 after

-1.77

POSITIVE LOGITS

↵↵

2.17

筇

2.11

啣

1.95

艄

1.92

 myriad

1.85

컹

1.84

“

1.84

鹩

1.84

 Infór

1.81

扨

1.81

Activations Density 0.013%