INDEX

Explanations

shell commands and variables

The neuron fires on shell‐script code tokens—things like the shebang (#!/…), variable assignments (VAR=…), array indexing (array[…]=…), loop keywords, etc.—i.e. Bash syntax elements.

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

 fuera

-1.47

水墨

-1.45

 außerhalb

-1.37

 zudem

-1.34

 simplemente

-1.29

 cAMP

-1.25

水彩

-1.23

 quotidienne

-1.23

 область

-1.22

白衣

-1.22

POSITIVE LOGITS

完成です

1.61

 kompati

1.52

 Tropfen

1.38

hopped

1.35

 становятся

1.35

menü

1.35

 аппети

1.34

 Coloque

1.34

쐈

1.32

vably

1.31

Activations Density 0.280%