INDEX

Explanations

the

The neuron activates on numeric tokens (both integers and decimals), flagging quantities whenever they appear.

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

 inspiradoras

-1.79

鯪

-1.77

Итог

-1.74

 consideradas

-1.66

 laquelle

-1.66

 москва

-1.66

 deň

-1.64

</h2>

-1.62

 století

-1.60

Bisous

-1.59

POSITIVE LOGITS

the

2.58

</i>

2.36

—

2.05

to

1.86

“

1.80

}\}

1.69

 bessere

1.68

 trusty

1.66

.”

1.64

1.63

Activations Density 0.013%