INDEX

Explanations

counting parts of a whole

The neuron fires on the word “of” when it’s used in numeric “X of Y” constructions (e.g. “1 of 2 Photos,” “Page 2 of 95,” etc.).

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

,”

-1.92

</b>

-1.73

谙

-1.69

UAWEI

-1.63

欵

-1.59

);

-1.59

樯

-1.55

if

-1.53

-1.46

 کفش

-1.46

POSITIVE LOGITS

 continuar

1.78

 ของ

1.70

 bekom

1.69

ים

1.66

Instead

1.62

 gese

1.61

 они

1.59

кото

1.57

还不错

1.55

 occasions

1.53

Activations Density 0.007%