INDEX

Explanations

common function words

This neuron detects numeric tokens (digits or number strings) in the text.

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

妧

-1.91

 réserver

-1.80

 trám

-1.77

 délais

-1.73

 encender

-1.72

 déposer

-1.66

 appeler

-1.63

 limpar

-1.63



-1.60

 sélectionner

-1.59

POSITIVE LOGITS

 with

1.83

or

1.80

 using

1.66

an

1.56

but

1.55

$\

1.47

1.46

to

1.41

 only

1.41

Activations Density 0.129%