INDEX

Explanations

question and answer format

This neuron detects structural/formatting tokens in documents—e.g. long runs of whitespace or indentation, email‐style headers and metadata markers.

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

 which

-1.52

if

-1.51

MARTIN

-1.36

 cleverly

-1.35

 after

-1.35

 revital

-1.33

 meticulously

-1.32

 where

-1.31

 Since

-1.30

 would

-1.27

POSITIVE LOGITS

 comprando

1.58

during

1.56

 acheter

1.45

bê

1.44

 fêtes

1.34

᠌

1.34

llten

1.33

🫅

1.30

they

1.30

槼

1.29

Activations Density 0.131%