INDEX

Explanations

programming data types

This neuron detects numeric literals (integers or 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

 حتی

-1.69

зык

-1.63

 hatta

-1.58

啬

-1.57

 เพราะ

-1.49

 بسبب

-1.47

}],

-1.45

 навіть

-1.38

 mnoho

-1.38

 многие

-1.38

POSITIVE LOGITS

 which

1.83

who

1.60

 тоже

1.56

 recupero

1.54

 whom

1.48

 cammino

1.41

 ebenfalls

1.38

 section

1.38

kých

1.36

 whose

1.35

Activations Density 0.025%