INDEX

Explanations

"the" followed by a noun

The neuron activates on numeric tokens—especially floating‐point numbers (e.g. “80,” “73.5,” “72.5,” etc.)—so it’s detecting occurrences of decimal‐format numbers.

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

</strong>

-1.47

 lyst

-1.32

uestran

-1.27

𝘣

-1.26

our

-1.26

 löytyy

-1.25

 przecież

-1.25

 cuando

-1.23

þe

-1.23

 both

-1.23

POSITIVE LOGITS

irmingham

1.40

 HAVING

1.34

少々

1.29

 lanza

1.27

そろそろ

1.27

ㆍ

1.25

 Encuentra

1.25

othermal

1.24

 ένα

1.23

1.20

Activations Density 0.069%