INDEX

Explanations

proteins, clubs, people, clothing, gardens

The neuron fires strongly on capitalized proper names and titles—i.e. multiword named entities such as brand names, organization names, place names, and other proper nouns.

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

and

-2.69

 with

-2.50

 also

-1.95

one

-1.82

 such

-1.81

for

-1.66

↵

-1.62

 based

-1.62

as

-1.52

 along

-1.52

POSITIVE LOGITS

 เป็น

1.67

 เว

1.60

 tiden

1.58

 huset

1.57

隹

1.55

actéristi

1.45

 ........

1.41

 heeft

1.41

 kommet

1.40

་་

1.38

Activations Density 0.267%