INDEX

Explanations

statement terminator

The neuron detects numeric literal tokens (in particular floating-point numbers) 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

 mereka

-1.41

 they

-1.40

viewDidLoad

-1.28

ulado

-1.25

 הם

-1.20

vanta

-1.18

 они

-1.17

urbation

-1.16

他们

-1.16

ornada

-1.16

POSITIVE LOGITS

its

1.60

1.58

 cantora

1.37

</h2>

1.36

1.34

 estadual

1.29

 brasile

1.27

 atriz

1.27

</h3>

1.25

''

1.25

Activations Density 0.021%