INDEX

Explanations

code structures and definitions

The neuron flags occurrences of the “.” member‐access operator (i.e. object.property or object.method calls).

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

is

-1.98

 also

-1.93

 about

-1.83

 just

-1.83

 even

-1.76

 such

-1.73

for

-1.70

by

-1.64

 along

-1.56

in

-1.55

POSITIVE LOGITS

 แต่

1.82

 quí

1.79

还不错

1.69

 некоторых

1.69

ཿ

1.68

开玩笑

1.67

휀

1.64

很赞

1.63

provides

1.59

 rús

1.58

Activations Density 0.034%