INDEX

Explanations

1 and 2

The neuron consistently fires on the standalone digit “2.”

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

.="

-1.39

为何

-1.32

-1.29

 بالفعل

-1.28

and

-1.28

-1.26

 бесплатно

-1.22

 tonel

-1.20

all

-1.18

 what

-1.18

POSITIVE LOGITS

are

1.91

 then

1.52

輜

1.41

École

1.40

凃

1.39

 dirancang

1.38

 Zähne

1.38

 olmak

1.35

 sécher

1.35

⃢

1.34

Activations Density 0.056%