INDEX

Explanations

programming code constructs

The neuron fires on capitalized programming identifiers—i.e. class or type names in 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

。

-1.55

任何

-1.38

and

-1.38

 europé

-1.37

-1.36

UDIO

-1.34

 ESPECIAL

-1.34

 Ebenso

-1.28

POSITIVE LOGITS

しまいます

1.33

elola

1.31

というわけで

1.25

 straips

1.24

can

1.24

+)$

1.24

 vaqueros

1.22

 }}">

1.20

 нашем

1.20

]}\

1.20

Activations Density 0.013%