INDEX

Explanations

The neuron fires on content‐bearing nouns—especially key objects or abstract concepts (e.g. products, philosophies, insights).

connecting two nouns

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 père

-1.55

 seront

-1.44

 boister

-1.40

 REALLY

-1.39

 ayant

-1.36

itself

-1.36

 dette

-1.35

regarding

-1.34

himself

-1.34

 VERY

-1.32

POSITIVE LOGITS

 which

1.68

 provide

1.62

 their

1.59

 plethora

1.57

 incredible

1.57

 enormous

1.56

 burgeoning

1.53

 activities

1.48

 giving

1.48

 putting

1.47

Activations Density 0.081%