INDEX

Explanations

illustrations and pictures

This neuron activates on common list-style separators and function words—especially commas, conjunctions (and), articles (of, the), and other high-frequency tokens that structure lists or itemizations.

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

 leaflets

-0.93

 during

-0.93

 manuscripts

-0.92

cellence

-0.89

or

-0.89

 članak

-0.87

createClass

-0.87

in

-0.86

 oktober

-0.85

 berichten

-0.84

POSITIVE LOGITS

 pictures

2.17

 illustrations

2.16

 diagrams

2.02

 graphics

1.95

 images

1.87

 photographs

1.83

 graphs

1.75

 photos

1.55

illustrations

1.44

 diagram

1.41

Activations Density 0.080%