INDEX

Explanations

allowed to, kit lenses, board game, line table, Freedom techniques, of life

The neuron fires on prominent topic‐or section‐heading keywords—i.e. major technical or domain–specific nouns that signal a new subject or heading.

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

 ως

-1.43

 них

-1.38

ódios

-1.34

 music

-1.31

why

-1.30

ystis

-1.30

ㄗ

-1.30

rões

-1.29

 uwag

-1.28

unyi

-1.27

POSITIVE LOGITS

1.45

but

1.41

any

1.37

 even

1.37

 from

1.24

чя

1.23

 both

1.23

 chale

1.17

 словно

1.16

 eten

1.14

Activations Density 0.217%