INDEX

Explanations

project directory paths

This neuron fires on numeric tokens—especially numbers embedded in file paths or code (e.g. path segments, file sizes, version 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

 almohada

-1.48

its

-1.47

 alleen

-1.42

 falda

-1.38

 تنها

-1.36

 tylko

-1.34

the

-1.34

 verdes

-1.28

 faites

-1.25

you

-1.24

POSITIVE LOGITS

Since

1.46

 from

1.46

fillment

1.42

 until

1.41

1.40

When

1.40

but

1.36

Our

1.35

Each

1.34

・・・」

1.34

Activations Density 0.089%