INDEX

Explanations

table separators and structure

This neuron detects table elements—especially pipes (“|”), separator dashes (“---”), and numeric cell entries—in formatted (e.g. Markdown‐style) tables.

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

桄

-1.34

⦊

-1.34

the

-1.30

 یکی

-1.30

相比

-1.24

其它

-1.24

nemen

-1.23

 教程

-1.23

gegen

-1.22

 anderen

-1.20

POSITIVE LOGITS

logio

1.27

 obligatorio

1.20

crom

1.19

 deseado

1.15

 obligado

1.15

EDES

1.13

 fijo

1.12

𝜋

1.11

 mundos

1.11

cester

1.10

Activations Density 0.022%