INDEX

Explanations

fields of knowledge and education

The neuron detects words naming broad topical domains or fields (e.g. “international,” “development,” “relief,” “universities,” “health”).

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

for

-1.82

 compared

-1.61

 regarding

-1.59

or

-1.55

 except

-1.54

 while

-1.51

 napisał

-1.46

solicited

-1.41

 that

-1.41

niño

-1.41

POSITIVE LOGITS

 нашей

1.43

 succes

1.41

 derselben

1.41

 seulement

1.36

very

1.35

 gitti

1.35

phur

1.34

 שלנו

1.30

tovers

1.29

Activations Density 0.065%