INDEX

Explanations

dynamic memory allocation

The neuron selectively fires on numeric literals (integer constants) 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

 certeza

-0.77

 habitu

-0.75

wydd

-0.75

 gekenn

-0.73

 nhan

-0.73

телям

-0.70

 hyö

-0.70

 gekennzeichnet

-0.69

 ממש

-0.69

選挙

-0.69

POSITIVE LOGITS

 malloc

1.95

malloc

1.92

 allocation

1.53

 allocating

1.50

ALLOC

1.50

 allocate

1.41

 Allocate

1.40

allocate

1.38

Allocate

1.38

alloc

1.37

Activations Density 0.027%