INDEX

Explanations

boolean conditions and properties

The neuron spots programming identifiers and constants (e.g. class names, enum variants, flags) 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

and

-2.17

but

-1.78

 včetně

-1.59

 during

-1.53

 would

-1.53

to

-1.48

ecco

-1.46

 that

-1.43

 with

-1.34

 possibly

-1.34

POSITIVE LOGITS

 disait

1.57

BindingResult

1.56

They

1.48

ズミ

1.47

Product

1.42

1.39

<bos>

1.39

 больше

1.38

www

1.37

Activations Density 0.088%