INDEX

Explanations

U.S. or United States

The neuron activates on the “U.” and “S.” tokens when they appear together as the U.S. abbreviation, i.e. it detects references to “U.S.” (United States).

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

 verhind

-1.98

戋

-1.72

 troisième

-1.70

 petites

-1.64

盉

-1.62

 ausein

-1.60

ಱ

-1.55

 erneu

-1.52



-1.52

quê

-1.49

POSITIVE LOGITS

of

2.13

 even

2.08

and

2.05

 because

2.02

 when

1.65

 there

1.63

as

1.60

 only

1.59

 Since

1.56

any

1.54

Activations Density 0.132%