INDEX

Explanations

descriptive states or qualities

The neuron strongly activates on common function words—especially short prepositions and articles like “of,” “by,” “except,” and “with.”

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

 faciliter

-1.33

まります

-1.27

 indispens

-1.27

瞟

-1.23

 конкурс

-1.23

 formid

-1.21

וּ

-1.17

 incredi

-1.16

 конкурса

-1.14

 جدید

-1.13

POSITIVE LOGITS

 this

1.59

 different

1.57

 things

1.55

 good

1.48

 some

1.46

 what

1.41

 enormous

1.40

 something

1.39

 everything

1.38

 both

1.33

Activations Density 0.071%