INDEX

Explanations

dates and times

The neuron chiefly fires on numeric tokens—particularly digits and multi‐digit numbers used in dates or identifiers.

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

 only

-1.86

 these

-1.71

two

-1.71

for

-1.65

 this

-1.60

 several

-1.60

 could

-1.59

 what

-1.56

<bos>

-1.53

any

-1.52

POSITIVE LOGITS

ようになります

1.54

場合には

1.54

 sauver

1.42

際には

1.41

あー

1.39

だけではなく

1.39

っているので

1.36

nalité

1.34

 représenter

1.33

 enfermos

1.32

Activations Density 0.078%