INDEX

Explanations

javascript code snippets

This neuron detects numeric literal values (number 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

of

-1.88

 sonre

-1.59

 Schließlich

-1.55

-1.45

 Darüber

-1.40

 Ведь

-1.38

 nacer

-1.37

 เพราะ

-1.36

 Даже

-1.35

杙

-1.35

POSITIVE LOGITS

ků

1.74

or

1.58



1.51

criptive

1.47

 aéreas

1.46

 adesivo

1.45

 hilsen

1.45

有個

1.44

umerable

1.43

しくは

1.43

Activations Density 0.013%